Open-Vocabulary Object Detection (OVOD) aims to detect both known and novel categories in complex visual scenes, surpassing the limitations of conventional closed-set detectors. Recent advances in vision-language models (VLMs) like CLIP have enabled zero-shot recognition by aligning visual features with large-scale textual embeddings. However, current OVOD approaches often fall short by overlooking critical contextual and semantic cues necessary for discovering a broader range of novel objects. To address this, we propose BFDet, a scene-to-object reasoning framework that leverages the complementary strengths of Large Language Models (LLMs) and VLMs. BFDet introduces a novel scene-to-object reasoning mechanism grounded in foreground-background context interaction. It first uses high-confidence objects to infer the scene-level background. This scene background then guides the discovery of foreground objects by prompting an LLM to generate scene-sensitive novel object candidates. These candidates are subsequently verified through cross-modal alignment and used as high-quality pseudo-labels to enrich detector training. Designed as a plug-and-play module, BFDet integrates seamlessly into existing detection pipelines and consistently improves performance on novel categories across COCO and LVIS benchmarks.

Time series anomaly detection (TSAD) is critical in various real-world applications. Due to the high cost of manual annotation, unsupervised methods are commonly employed to distinguish abnormal patterns from normal ones based on data or representation characteristics. However, the limited coverage of a single dataset often leads to misclassifying test-time normal patterns that deviate from the training distribution as anomalies. In view of this, we propose to introduce domain knowledge from auxiliary datasets (AuxSets) to enhance domain-level normality understanding in the target dataset (TargetSet). However, through in-depth analysis on the representation space of the TargetSet after incorporating AuxSets, we find that consistent knowledge about normality from homogeneous AuxSets do little help to TargetSet, while diverse knowledge from heterogeneous AuxSets can bring semantic confusion of normality for TargetSet, both of which can degrade TargetSet detection performance. To address the issue, we design DoKnowAD, a framework that introduces a Representation HyperVolume Estimation metric to identify helpful heterogeneous AuxSets, and further adopts contrastive learning to enforce loose coupling between datasets and high cohesion within single dataset to calibrate the TargetSet’s representation space, thus mitigating knowledge confusion. Extensive experiments on five popular datasets across different domains demonstrate that DoKnowAD consistently outperforms existing TSAD baselines in various metrics.

The Job-Shop Scheduling Problem (JSSP), under various forms of manufacturing uncertainty, has recently attracted considerable research attention. Most existing studies focus on parameter uncertainty, such as variable processing times, and typically adopt the actor-critic framework. In this paper, we explore a different but prevalent form of uncertainty in JSSP: structural uncertainty. Structural uncertainty arises when a job may follow one of several routing paths, and the selection is determined not by policy, but by situational factors (e.g., the quality of intermediate products) that cannot be known in advance. Existing methods struggle to address this challenge due to incorrect credit assignment: a high-quality action may be unfairly penalized if it is followed by a time-consuming path. To address this problem, we propose a novel method named UP-AAC. In contrast to conventional actor-critic methods, UP-AAC employs an asymmetric architecture. While its actor receives a standard stochastic state, the critic is crucially provided with a deterministic state reconstructed in hindsight. This design allows the critic to learn a more accurate value function, which in turn provides a lower-variance policy gradient to the actor, leading to more stable learning. In addition, we design an attention-based Uncertainty Perception Model (UPM) to enhance the actor's scheduling decisions. Extensive experiments demonstrate that our method outperforms existing approaches in reducing makespan on benchmark instances.

Recent breakthroughs in Transformer-based large models, have driven widespread tasks, yet their reliance on centralized cloud deployment raises significant privacy risks due to sensitive data exposure.While edgebased collaborative inference offers a privacypreserving alternative, existing methods face critical limitations: static model partitioning cannot adapt to dynamic edge resource fluctuations, and rigid multi-head attention handling overlooks semantic-critical prioritization and parallelism.We propose EdgeFormer, a latency-aware framework for distributed Transformer inference in resource-constrained edge networks.EdgeFormer dynamically allocates model blocks across devices via efficiencystorage trade-off optimization and introduces collaborative Multi-Head Attention (cMHA), which distributes semantic-critical attention heads across devices while pruning redundant ones under real-time constraints.We further develop LiScore, a composite metric integrating attention diversity and latency costs, alongside a similarity-based retrieval method to reduce recomputation overhead.Extensive experiments demonstrate that EdgeFormer achieves up to 2.01× inference acceleration over state-of-theart baselines with ≤1.06% accuracy loss, maintaining robustness under varying edge conditions.

This article presents a model predictive control (MPC) method for humanoid dual-arm robots (DARs) to realize hybrid force-velocity control in the task space. The proposed method, named hybrid force-velocity MPC (HFV-MPC), enables simultaneous tracking of velocity and external force in the task space, as well as internal force regulation. First, a synchronous decoupling model of internal force, external force, and velocity is developed to address the limitations of conventional methods that decouple internal and external forces solely in the end-effector space and fail to meet the independent regulation requirements of external force and velocity in the task space. Furthermore, a generalized velocity model that integrates task-space velocity, external force, and internal force was established, and a HFV-MPC framework is constructed to enhance system robustness and achieve precise dual-arm force–velocity tracking. Subsequently, a comprehensive MPC cost function is designed, incorporating a dual-arm motion coordination coefficient to suppress undesired redundant joint motions and improve whole-body motion stability. In addition, a feedforward-linearized incremental system is derived, where the incremental model and force prediction jointly construct a generalized state-space model for MPC, to balance modeling accuracy and computational efficiency. Finally, the proposed method is validated in terms of effectiveness and robustness through both simulation and physical robotic experiments.

Due to the selective absorption and scattering of light by diverse aquatic media, underwater images usually suffer from various visual degradations. Existing underwater image enhancement (UIE) approaches that combine underwater physical imaging models with neural networks often fail to accurately estimate imaging model parameters such as scene depth and veiling light, resulting in poor performance in certain scenarios. To address this issue, we propose a physical model-guided framework for jointly training a Deep Degradation Model (DDM) with any advanced UIE model. DDM includes three well-designed sub-networks to accurately estimate various imaging parameters: a veiling light estimation sub-network, a factors estimation sub-network, and a depth estimation sub-network. Based on the estimated parameters and the underwater physical imaging model, we impose physical constraints on the enhancement process by modeling the relationship between underwater images and desired clean images, i.e., outputs of the UIE model. Moreover, while our framework is compatible with any UIE model, we design a simple yet effective fully convolutional UIE model, termed UIEConv. UIEConv utilizes both global and local features through a dual-branch structure. UIEConv trained within our framework achieves remarkable enhancement results across diverse underwater scenes. Furthermore, as a byproduct of UIE, the trained depth estimation sub-network enables accurate underwater scene depth estimation. Extensive experiments conducted in various real underwater imaging scenarios, including deep-sea environments with artificial light sources, validate the effectiveness of our framework and the UIEConv model. Code is available at https://github.com/ddz16/UWEnhancer.

With the rapid development of digital technologies, a large range of real-world systems, spanning from cloud servers, IoT devices, to industrial control systems, continuously generate vast amounts of time series data. Time series anomaly detection (AD) plays a crucial role in maintaining system stability by identifying unusual patterns from normal distributions, with the primary challenge lies in learning effective anomaly-discriminative representations. Recently, diffusion models have been applied to time series AD due to their strong representational capabilities. However, existing diffusion-based methods typically rely on reconstruction errors, which not only fail to fully exploit the representational potential of diffusion models but also be computationally intensive. To address these limitations, through experimental observation and theoretical analysis, we show thatspecific regions of the diffusion noises exhibit stronger representation capabilitiesfor normal patterns, which can be leveraged to enhance AD performance and reduce computational costs. Building on these insights, we propose NoiseAD, a diffusion noise-guided anomaly detection method incorporating an optimal noise steps selection approach to identify diffusion steps with higher resolution. Extensive experiments on diverse benchmarks demonstrate the superiority of NoiseAD over state-of-the-art methods, further substantiated by insightful visualizations. Code could be available athttps://github.com/shiwang-Xing/NoiseAD.

Dense video captioning (DVC) aims to describe multiple events within a video, and its performance is greatly affected by the accuracy of video event detection. Video event detection involves predicting the proposal boundaries (start and end times) and the classification score of each event in a video. Recently, a few methods have applied diffusion models originally designed for image object detection to detect events in DVC. These methods add noise to the ground-truth event proposal boundaries, and subsequently learn the denoising process. However, these methods often overlook the fundamental differences between videos and images. We observe that, whereas in images the important information for object classification is normally around the boundaries of the ground-truth boxes, in videos the key information for event classification is typically centered in the middle of ground-truth event proposals. As a result, the classification module in these existing diffusion models becomes insensitive to boundary changes introduced by the added noise, leading to sub-optimal performance. This paper introduces DiffDVC, an innovative diffusion model for DVC. The core of DiffDVC is a boundary-sensitive detector. The detector increases the sensitivity of the classification module to boundary changes by focusing on frames within a specific range around the start and end times of noisy event proposals. Additionally, this range is dynamically adjusted to suit different event proposals. Comprehensive experiments on ActivityNet-1.3, ActivityNet Captions, and YouCook2 datasets show DiffDVC achieving superior performance.

Byte Pair Encoding (BPE) serves as a foundation method for text tokenization in the Natural Language Processing (NLP) field. Despite its wide adoption, the original BPE algorithm harbors an inherent flaw: it inadvertently introduces a frequency imbalance for tokens in the text corpus. Since BPE iteratively merges the most frequent token pair in the text corpus to generate a new token and keeps all generated tokens in the vocabulary, it unavoidably holds tokens that primarily act as components of a longer token and appear infrequently on their own. We term such tokens as Scaffold Tokens. Due to their infrequent occurrences in the text corpus, Scaffold Tokens pose a learning imbalance issue. To address that issue, we propose Scaffold-BPE, which incorporates a dynamic scaffold token removal mechanism by parameter-free, computation-light, and easy-to-implement modifications to the original BPE method. This novel approach ensures the exclusion of low-frequency Scaffold Tokens from the token representations for given texts, thereby mitigating the issue of frequency imbalance and facilitating model training. On extensive experiments across language modeling and even machine translation, Scaffold-BPE consistently outperforms the original BPE, well demonstrating its effectiveness.

Recently, Large language models (LLMs) have revolutionized Natural Language Processing (NLP). Pretrained LLMs, due to limited training context size, struggle with handling long token sequences, limiting their performance on various downstream tasks. Current solutions toward long context modeling often employ multi-stage continual pertaining, which progressively increases the effective context length through several continual pretraining stages. However, those approaches require extensive manual tuning and human expertise. In this paper, we introduce a novel single-stage continual pretraining method, Head-Adaptive Rotary Position Embedding (HARPE), to equip LLMs with long context modeling capabilities while simplifying the training process. Our HARPE leverages different Rotary Position Embedding (RoPE) base frequency values across different attention heads and directly trains LLMs on the target context length. Extensive experiments on 4 language modeling benchmarks, including the latest RULER benchmark, demonstrate that HARPE excels in understanding and integrating long-context tasks with single-stage training, matching and even outperforming existing multi-stage methods. Our results highlight that HARPE successfully breaks the stage barrier for training LLMs with long context modeling capabilities.

Yizhe Xiong, Xiansheng Chen, Xin Ye, Hui Chen, Zijia Lin, Haoran Lian, Zhenpeng Su, Wei Huang, Jianwei Niu, Jungong Han, Guiguang Ding. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

The prevalent use of Byte Pair Encoding (BPE) in Large Language Models (LLMs) facilitates robust handling of subword units and avoids issues of out-of-vocabulary words. Despite its success, a critical challenge persists: long tokens, rich in semantic information, have fewer occurrences in tokenized datasets compared to short tokens, which can result in imbalanced learning issue across different tokens. To address that, we propose LBPE, which prioritizes long tokens during the encoding process. LBPE generates tokens according to their descending order of token length rather than their ranks in the vocabulary, granting longer tokens higher priority during the encoding process. Consequently, LBPE smooths the frequency differences between short and long tokens, and thus mitigates the learning imbalance. Extensive experiments across diverse language modeling tasks demonstrate that LBPE consistently outperforms the original BPE, well demonstrating its effectiveness.

In packet erasure networks, federated learning (FL) typically suffers more prohibitive communication overhead from massive retransmissions of high-dimensional gradients. As a result, recent studies are dedicated to developing retransmission-free gradient compression techniques with erasure resilience. Nonetheless, two limitations remain unsolved: existing works neither explore why packet erasure degrades the performance of FL nor exploit the spatial correlations among gradient entries for better compression. In this paper, we investigate FL performance degradation via analyzing model updating deviation and find that the deviation is exacerbated by dependencies among lost gradient entries. On top of this observation, we propose FedRIVQ, a communication-efficient and robust FL framework taking a customized compressor termed random interleaved vector quantization (VQ). FedRIVQ leverages the spatial correlations among gradient entries with VQ and randomly interleaves these entries prior to VQ to eliminate their dependencies. These innovations allow all gradient entries to share an identical erasure probability, thereby packet erasure is equivalent to random erasure, which significantly improves both communication efficiency and the robustness of FL. Theoretical analysis and experimental results consistently demonstrate the effectiveness of our designs.

The out-of-distribution (OOD) generalization problem in federated learning (FL) has recently attracted significant research interest. A common approach, derived from centralized learning, is to extract causal features which exhibit causal relationships with the label. However, in FL, the global feature extractor typically captures only invariant causal features shared across clients and thus discards many other causal features that are potentially useful for OOD generalization. To address this problem, we propose FedUni, a simple yet effective architecture trained to extract all possible causal features from any input. FedUni consists of a comprehensive feature extractor, designed to identify a union of all causal feature types in the input, followed by a feature compressor, which discards potential \textit{inactive} causal features. With this architecture, FedUni can benefit from collaborative training in FL while avoiding the cost of model aggregation (i.e., extracting only invariant features). In addition, to further enhance the feature extractor's ability to capture causal features, FedUni add a causal intervention module on the client side, which employs a counterfactual generator to generate counterfactual examples that simulate distributions shifts. Extensive experiments and theoretical analysis demonstrate that our method significantly improves OOD generalization performance.

Robots powered by large language model (LLM) demonstrate significant research and application potential by effectively interpreting scene information to respond to human commands. However, when robots rely on static scene information during task execution, they face difficulties in adapting to changes in the environment, posing a major challenge for dynamic scene perception. To address the above issues, we propose an innovative interaction-driven approach to enhance robots' ability to perceive dynamic scene information. This approach consists of two contributions, the observation point selection module and the dynamic scene maintenance module. Specifically, first, the robot uses the 3D scene graph (3DSG) containing assets and objects to perceive static scene information through the LLM planner. Next, the best observation point for each asset is obtained through the observation point selection module. Then, with the help of the best observation point, the dynamic scene maintenance module interacts with the asset-related objects to dynamically update all the object node information related to the asset node. This approach enables robots to maintain dynamic scene information, enhancing their adaptability in unpredictable environments and improving task reliability. We evaluated our method using the iTHOR and RoboTHOR datasets within the AI2-THOR simulator and in real-world scenarios. Experimental results demonstrate that our method effectively and accurately maintains robots' perception of dynamic scene information.

Effective information interaction can enhance the coordination capabilities in collaborative multi-agent reinforcement learning (MARL). A popular communication scheme is the exchange of agents’ intention information, which typically involves broadcasting agents’ intention information to all other agents. This not only increases the communication overhead of the entire system but also interferes with the decision-making of agents to some extent due to the reception of irrelevant intention information from other agents. In this paper, we propose the Multi-Agent Intention Filter (MAIF), which filters out intentions irrelevant to the agent, allowing the agent to focus on the intentions of agents with whom it is more likely to collaborate, thus promoting effective cooperation among agents. Specifically, our method first uses a State Simulator to coordinate the joint intentions of agents based on their current joint observations and predict target states, providing prior knowledge for subsequent intention filtering. Then, we calculate the causal effect of agents’ intentions on target states and filter out intentions that are irrelevant to the agent, thereby promoting effective cooperation. Each agent will use the filtered intention information to assist in decision-making. Experimental results show that our method outperforms strong baselines in multiple cooperative MARL tasks under various task settings.

Dense video captioning aims to generate descriptive sentences for each temporally localized event in a video. This task comprises two subtasks: event detection and event captioning. Existing methods commonly adopt a DETR-like (Detection Transformer) architecture to perform both subtasks in parallel. These methods assume that both subtasks require the same visual information and thus extract a single event representation for each event using a shared query. We observe that event detection and event captioning emphasize different regions of a video. In particular, compared to event captioning, event detection tends to focus more on the boundary regions of event proposals. Therefore, relying on shared queries may hinder the ability of the model to meet the specific needs of each subtask, leading to suboptimal performance. In this paper, we propose decoupling the two subtasks by assigning distinct queries to each, enabling more accurate capture of task-specific features. Specifically, we introduce a task-specific query transformation module. This module utilizes two sets of task-specific prompts to transform shared queries into queries tailored for each subtask. These task-specific queries enable each subtask to attend to the video regions that are most beneficial to its respective objectives. By integrating our method into several state-of-the-art frameworks, we achieve superior performance on both event detection and event captioning.

Federated Learning (FL) faces challenges due to data heterogeneity, which limits the global model’s performance across diverse client distributions. Personalized Federated Learning (PFL) addresses this by enabling each client to process an individual model adapted to its local distribution. Many existing methods assume that certain global model parameters are difficult to train effectively in a collaborative manner under heterogeneous data. Consequently, they localize or fine-tune these parameters to obtain personalized models. In this paper, we reveal that both the feature extractor and classifier of the global model are inherently strong, and the primary cause of its suboptimal performance is the mismatch between local features and the global classifier. Although existing methods alleviate this mismatch to some extent and improve performance, we find that they either (1) fail to fully resolve the mismatch while degrading the feature extractor, or (2) address the mismatch only post-training, allowing it to persist during training. This increases inter-client gradient divergence, hinders model aggregation, and ultimately leaves the feature extractor suboptimal for client data. To address this issue, we propose FedPFT, a novel framework that resolves the mismatch during training using personalized prompts. These prompts, along with local features, are processed by a shared self-attention-based transformation module, ensuring alignment with the global classifier. Additionally, this prompt-driven approach offers strong flexibility, enabling task-specific prompts to incorporate additional training objectives (\eg, contrastive learning) to further enhance the feature extractor. Extensive experiments show that FedPFT outperforms state-of-the-art methods by up to 5.07%, with further gains of up to 7.08% when collaborative contrastive learning is incorporated.

Mobile edge computing (MEC), along with device-to-device (D2D) assisted MEC (D-MEC), are promising technologies that could improve the quality-of-experience for mobile devices (MDs) by offloading their tasks to edge servers or nearby idle MDs. There is a popular trend to develop distributed task offloading algorithms using multi-agent reinforcement learning (MARL), whose adoption of central critics during training makes the offloading still size-sensitive. Therefore, this paper proposes a Size-Insensitive Task Offloading (SITOff) algorithm for D-MEC based on fully-distributed offloading without maintaining any central venue. Specifically, taking advantage of the inherent graph-like structure of D-MEC, SITOff adopts graphs to represent MDs’ states and relationships and form each MD's local knowledge about D-MEC through graph computation. Furthermore, considering the limitation of local knowledge in performing whole performance-oriented offloading, each MD utilizes D2D-transmitting to exchange knowledge with its neighbors and form a comprehensive knowledge about D-MEC to enhance the coordination of distributed offloading. Additionally, regarding the different impacts of neighbors’ knowledge, each MD leverages attention mechanisms to selectively learn its neighbors’ knowledge during knowledge-exchange. Extensive experimental results show the superiority of SITOff over state-of-the-art MARL-based offloading algorithms in D-MEC with various MDs, and the easy collaboration of SITOff with curriculum-learning for large-scale D-MEC offloading.

Personalized federated learning (PFL) allows clients to collaboratively train their personalized models to handle situations where data from different clients are not independent and identically distributed (non-IID). Previous PFL research implicitly assumes that clients benefit most from those with similar data distributions. Correspondingly, methods such as personalized weight aggregation assign higher weights to similar clients during aggregation. We pose a question: can a client benefit from other clients with dissimilar data distributions, and if so, how? This question is particularly relevant in scenarios with a high degree of non-IID, where clients have widely different distributions, and learning from only similar clients will result in a loss of knowledge from many other clients. We note that when dealing with clients with similar distributions, current methods tend to enforce their models to be close in the parameter space. It is reasonable to conjecture that a client can benefit from dissimilar clients if we allow their models to depart from each other. Based on this idea, we propose DiversiFed, which allows each client to learn from clients with diversified distribution. DiversiFed pushes personalized models of clients with dissimilar distributions apart in the parameter space while pulling together those with similar distributions. In addition, to achieve the above effect without using prior knowledge of distribution, we design a loss function that leverages model similarity to determine the degree of attraction and repulsion between any two models. Experiments on benchmark and medical datasets show that DiversiFed can outperform the state-of-the-art (SOTA) methods by up to 3.19%.

Personalized federated learning (PFL) is a popular distributed learning framework that allows clients to have different models and has many applications where clients' data are in different domains, including autonomous driving, traffic surveillance, and medical diagnosis. The typical model of a client in PFL features a global encoder trained by all clients to extract universal features from the raw data and personalized layers (e.g., a classifier) trained using the client's local data. Nonetheless, due to the differences between the data distributions of different clients (also known as, domain gaps), the universal features produced by the global encoder largely encompass numerous components irrelevant to a certain client's local task. Some recent PFL methods address the above problem by personalizing specific parameters within the encoder. However, these methods encounter substantial challenges attributed to the high dimensionality and nonlinearity of neural network parameter space. In contrast, the feature space exhibits a lower dimensionality, providing greater intuitiveness and interpretability as compared to the parameter space. To this end, we propose a novel PFL framework named FedPick. FedPick achieves PFL within the low-dimensional feature space by adaptively selecting task-relevant features for each client from the features generated by the global encoder based on its local data distribution. It presents a more accessible and interpretable implementation of PFL compared to those methods working in the parameter space. Extensive experimental results on multiple cross-domain datasets show that FedPick can effectively select task-relevant features for each client and improve model performance in cross-domain FL.

We propose 3DFaceSculptor, a general-purpose framework for interactive 3D face editing. Given a source 3D face mesh with semantic materials, and a user-specified semantic image, 3DFaceSculptor can accurately edit the source mesh following the shape guidance of the semantic image, while preserving the source topology as rigid as possible. Recent studies on generating 3D faces focus on learning neural networks to predict 3D shapes, which requires high-cost 3D training datasets. These learning-based methods are limited in compatibility and can only handle face styles involved in the training datasets. Unlike these methods, our 3DFaceSculptor is a non-training and common framework, which only requires supervision from readily-available semantic images, and is compatible with producing various face styles unlimited by datasets. In 3DFaceSculptor, based on the differentiable renderer technique, we deform the source face mesh according to the correspondences between semantic images and mesh materials. However, guiding complex 3D shapes with a simple 2D image incurs extra challenges, that is, the deformation accuracy, surface smoothness, geometric rigidity, and global synchronization of the edited mesh must be guaranteed. To address these challenges, we propose a hierarchical optimization architecture to balance the global and local shape features, and further propose various strategies and losses to improve properties of accuracy, smoothness, rigidity, and so on. Extensive experiments show that our 3DFaceSculptor is able to produce impressive results and has reached the state-of-the-art level.

Network-wide heavy hitter detection is usually performed by sampling on several network measurement points (NMPs) and merging the measurement results in the centralized controller to get a network-wide view. However, a packet may pass several NMPs and be counted multiple times when measurement results are merged, which causes the double-counting problem and leads to incorrect detection. Existing studies either overlook this problem or require significant memory usage. This paper proposes SAROS, a self-adaptive routing oblivious sampling method for accurate network-wide heavy hitter detection. Specifically, SAROS exploits a sampling mechanism in the data plane, where the sampling threshold on each measurement point is predicted and adaptively set by the control plane. Such guidance from the control plane greatly reduces the memory usage in the data plane, while mitigating the double-counting problem. Experimental results show that, compared with existing solutions, SAROS improves the F1-Score of heavy hitter detection by 10 ∼ 40%.

Document-level relation extraction (DocRE)aims to identify relationships between entities within a document.Due to the vast number of entity pairs, fully annotating all fact triplets is challenging, resulting in datasets with numerous false negative samples.Recently, selftraining-based methods have been introduced to address this issue.However, these methods are purely black-box and sub-symbolic, making them difficult to interpret and prone to overlooking symbolic interdependencies between relations.To remedy this deficiency, our insight is that symbolic knowledge, such as logical rules, can be used as diagnostic tools to identify conflicts between pseudo-labels.By resolving these conflicts through logical diagnoses, we can correct erroneous pseudolabels, thus enhancing the training of neural models.To achieve this, we propose Log-icST, a neural-logic self-training framework that iteratively resolves conflicts and constructs the minimal diagnostic set for updating models.Extensive experiments demonstrate that LogicST significantly improves performance and outperforms previous state-of-the-art methods.For instance, LogicST achieves an increase of 7.94% in F1 score compared to CAST (Tan et al., 2023a) on the DocRED benchmark (Yao et al., 2019).Additionally, LogicST is more time-efficient than its self-training counterparts, requiring only 10% of the training time of CAST.Code is available at https: //github.com/XingYing-stack/LogicST.

Federated learning (FL) generally suffers significant communication overhead from high-traffic gradient synchronization. The majority of existing studies on this problem aim at compressing gradients under the premise of reliable transmission. While transmission reliability can be ensured via TCP by default, the notably increased latency and retransmitted packets are prohibitive for most clients in FL. To tackle this issue, we propose FedMDC, a retransmission-free compression framework for FL over packet lossy networks. Given clients’ limited resources, FedMDC adopts multiple description coding to encode gradients into redundant descriptions for erasure resilience simply through multiplying an overcomplete matrix; and then quantizes these descriptions for compression. To further reduce quantization distortion and computational overhead, a reduced decoding algorithm is developed by decoding the aggregation of all clients’ encodings in conjunction with a customized dither quantization design. Besides, FedMDC explicitly supports adaptive bitrates subject to clients’ heterogeneous communication budgets, which maximize resource utilization to facilitate distortion reduction and accelerate model convergence. Theoretical analysis and experimental results both demonstrate the effectiveness of our scheme.

Neural implicit representations have been explored to enhance visual SLAM algorithms, especially in providing high-fidelity dense map. Existing methods operate robustly in static scenes but struggle with the disruption caused by moving objects. In this paper we present NID-SLAM, which significantly improves the performance of neural SLAM in dynamic environments. We propose a new approach to enhance inaccurate regions in semantic masks, particularly in marginal areas. Utilizing the geometric information present in depth images, this method enables accurate removal of dynamic objects, thereby reducing the probability of camera drift. Additionally, we introduce a keyframe selection strategy for dynamic scenes, which enhances camera tracking robustness against large-scale objects and improves the efficiency of mapping. Experiments on publicly available RGB-D datasets demonstrate that our method outperforms competitive neural SLAM approaches in tracking accuracy and mapping quality in dynamic environments.

Computation offloading is one of the key issues in mobile edge computing (MEC) that alleviates the tension between user equipment's limited capabilities and mobile application's high requirements. To achieve model-free computation offloading when reliable MEC dynamics are unavailable, deep reinforcement learning (DRL) has become a popular methodology. However, most existing DRL-based offloading approaches are developed for a single MEC environment, with invariant system bandwidth, edge capability, task types, etc., while realistic MEC scenarios tend to be of high diversity. Unfortunately, in multi-MEC environments, DRL-based offloading faces at least two challenges, learning inefficiency and interference of offloading experiences. To address the challenges, we propose a DRL-based Multi-environmental Module-compositional Modelfree computation OFFloading (M3OFF) framework. M3OFF generates offloading policies using module composition instead of a single DRL network so that learning efficiency could be improved by reusing the same modules and learning interference could be reduced by composing different modules. Furthermore, we design multiple module composition-specific training methods for M3OFF, including alternate modules-and-composer updates to improve training stability, loss-regularization to avoid module degeneration, and module-dropout to mitigate overfitting. Extensive experimental results on both simulation and testbed demonstrate that M3OFF outperforms the performances of most state-of-the-arts in multi-MEC and reaches close to single-MEC.

Federated learning (FL) enables distributed training via periodically synchronizing model updates among participants. Communication overhead becomes a dominant constraint of FL since participating clients usually suffer from limited bandwidth. To tackle this issue, top-k based gradient compression techniques are broadly explored in FL context, manifesting powerful capabilities in reducing gradient volumes via picking significant entries. However, previous studies are primarily conducted on the raw gradients where massive spatial redundancies exist and positions of non-zero (top-k) entries vary greatly between gradients, which both impede the achievement of deeper compressions. Top-k may also degrade the performance of trained models due to biased gradient estimations. Targeting the above issues, we propose FedTC, a novel transform coding based compression framework. FedTC transforms gradients into a new domain with more compact energy distributions, which facilitates reducing spatial redundancies and biases in subsequent sparsification. Furthermore, non-zero entries across clients from different rounds become highly aligned in the transform domain, motivating us to partition the gradients into smaller entry blocks with various alignment levels to better exploit these alignments. Lastly, positions and values of non-zero entries are independently compressed in a block-wise manner with our customized designs, through which a higher compression ratio is achieved. Theoretical analysis and extensive experiments consistently demonstrate the effectiveness of our approach.

Deep learning models often suffer performance degradation when test data diverges from training data. Test-Time Adaptation (TTA) aims to adapt a trained model to the test data distribution using unlabeled test data streams. In many real-world applications, it is quite common for the trained model to be deployed across multiple devices simultaneously. Although each device can execute TTA independently, it fails to leverage information from the test data of other devices. To address this problem, we introduce Federated Learning (FL) to TTA to facilitate on-the-fly collaboration among devices during test time. The workflow involves clients (i.e., the devices) executing TTA locally, uploading their updated models to a central server for aggregation, and downloading the aggregated model for inference. However, implementing FL in TTA presents many challenges, especially in establishing inter-client collaboration in dynamic environment, where the test data distribution on different clients changes over time in different manners. To tackle these challenges, we propose a server-side Temporal-Spatial Aggregation (TSA) method. TSA utilizes a temporal-spatial attention module to capture intra-client temporal correlations and inter-client spatial correlations. To further improve robustness against temporal-spatial heterogeneity, we propose a heterogeneity-aware augmentation method and optimize the module using a self-supervised approach. More importantly, TSA can be implemented as a plug-in to TTA methods in distributed environments. Experiments on multiple datasets demonstrate that TSA outperforms existing methods and exhibits robustness across various levels of heterogeneity. The code is available at https://github.com/ZhangJiayuan-BUAA/FedTSA.

To address data heterogeneity, the key strategy of Personalized Federated Learning (PFL) is to decouple general knowledge (shared among clients) and client-specific knowledge, as the latter can have a negative impact on collaboration if not removed. Existing PFL methods primarily adopt a parameter partitioning approach, where the parameters of a model are designated as one of two types: parameters shared with other clients to extract general knowledge and parameters retained locally to learn client-specific knowledge. However, as these two types of parameters are put together like a jigsaw puzzle into a single model during the training process, each parameter may simultaneously absorb both general and client-specific knowledge, thus struggling to separate the two types of knowledge effectively. In this paper, we introduce FedDecomp, a simple but effective PFL paradigm that employs parameter additive decomposition to address this issue. Instead of assigning each parameter of a model as either a shared or personalized one, FedDecomp decomposes each parameter into the sum of two parameters: a shared one and a personalized one, thus achieving a more thorough decoupling of shared and personalized knowledge compared to the parameter partitioning method. In addition, as we find that retaining local knowledge of specific clients requires much lower model capacity compared with general knowledge across all clients, we let the matrix containing personalized parameters be low rank during the training process. Moreover, a new alternating training strategy is proposed to further improve the performance. Experimental results across multiple datasets and varying degrees of data heterogeneity demonstrate that FedDecomp outperforms state-of-the-art methods up to 4.9%. The code is available at https://github.com/XinghaoWu/FedDecomp

In personalized federated learning (PFL), it is widely recognized that achieving both high model generalization and effective personalization poses a significant challenge due to their conflicting nature. As a result, existing PFL methods can only manage a trade-off between these two objectives. This raises an interesting question: Is it feasible to develop a model capable of achieving both objectives simultaneously? Our paper presents an affirmative answer, and the key lies in the observation that deep models inherently exhibit hierarchical architectures, which produce representations with various levels of generalization and personalization at different stages. A straightforward approach stemming from this observation is to select multiple representations from these layers and combine them to concurrently achieve generalization and personalization. However, the number of candidate representations is commonly huge, which makes this method infeasible due to high computational costs. To address this problem, we propose DualFed, a new method that can directly yield dual representations correspond to generalization and personalization respectively, thereby simplifying the optimization task. Specifically, DualFed inserts a personalized projection network between the encoder and classifier. The pre-projection representations are able to capture generalized information shareable across clients, and the post-projection representations are effective to capture task-specific information on local clients. This design minimizes the mutual interference between generalization and personalization, thereby achieving a win-win situation. Extensive experiments show that DualFed can outperform other FL methods. Code is available at https://github.com/GuogangZhu/DualFed.

LiDAR technology holds vast potential across various sectors, including robotics, autonomous driving, and urban planning. However, the performance of current LiDAR sensors is hindered by limited field of view (FOV), low resolution, and lack of flexible focusing capability. We introduce αLiDAR, an innovative LiDAR system that employs controllable actuation to provide a panoramic FOV, high resolution, and adaptable scanning focus. The core concept of αLiDAR is to expand the operational freedom of a LiDAR sensor through the incorporation of a controllable, active rotational mechanism. This modification allows the sensor to scan previously inaccessible blind spots and focus on specific areas of interest in an adaptive manner. By modeling uncertainties in LiDAR rotation process and estimating point-wise uncertainty, αLiDAR can correct point cloud distortions resulted from significant rotation. In addition, by optimizing LiDAR's rotation trajectory, αLiDAR can swiftly adapt to dynamic areas of interest. We developed several prototypes of αLiDAR and conducted comprehensive evaluations in various indoor and outdoor real-world scenarios. Our results demonstrate that αLiDAR achieves centimeter-level pose estimation accuracy, with an average latency of only 37 ms. In two typical LiDAR applications, αLiDAR significantly enhances 3D mapping accuracy, coverage, and density by 8.5×, 2×, and 1.6× respectively, compared to conventional LiDAR sensors. Additionally, αLiDAR's adaptive rotation improves the effective sensing distance by 1.8× and increases the number of perceived objects by 1.9×. A video demonstration of αLiDAR's in action in real world is available at https://youtu.be/x4zc_I_xTaw. The code is available at https://github.com/HViktorTsoi/alpha_lidar.

We present αLiDAR, an innovative LiDAR system that incorporates a controllable active rotational mechanism to broaden the field of view (FOV), enhance resolution, and provide adaptable focusing. This system addresses the inherent limitations of traditional LiDAR sensors, such as narrow FOV, low resolution, and lack of flexible focusing capability. By scanning blind spots and dynamically focusing on areas of interest, αLiDAR significantly surpasses conventional LiDAR sensors. Our prototypes, tested under varied real-world conditions, have demonstrated marked improvements in typical LiDAR applications. Specifically, αLiDAR enhances 3D mapping accuracy, coverage, and density by factors of 8.5, 2, and 1.6, respectively. Furthermore, the adaptive rotational mechanism of αLiDAR extends the effective sensing distance by 1.8× and increases object detection by 1.9×. To see αLiDAR in action, visit our video demonstration at https://youtu.be/x4zc_I_xTaw. Both the hardware and software implementations of αLiDAR are open-sourced at https://github.com/HViktorTsoi/alpha_lidar.

Federated learning is a distributed machine learning paradigm designed to protect user data privacy, which has been successfully implemented across various scenarios. In traditional federated learning, the entire parameter set of local models is updated and averaged in each training round. Although this full network update method maximizes knowledge acquisition and sharing for each model layer, it prevents the layers of the global model from cooperating effectively to complete the tasks of each client, a challenge we refer to as layer mismatch. This mismatch problem recurs after every parameter averaging, consequently slowing down model convergence and degrading overall performance. To address the layer mismatch issue, we introduce the FedPart method, which restricts model updates to either a single layer or a few layers during each communication round. Furthermore, to maintain the efficiency of knowledge acquisition and sharing, we develop several strategies to select trainable layers in each round, including sequential updating and multi-round cycle training. Through both theoretical analysis and experiments, our findings demonstrate that the FedPart method significantly surpasses conventional full network update strategies in terms of convergence speed and accuracy, while also reducing communication and computational overheads.

Manga is a fashionable Japanese-style comic form that is composed of black-and-white strokes and is generally displayed as raster images on digital devices. Typical mangas have simple textures, wide lines, and few color gradients, which are vectorizable natures to enjoy the merits of vector graphics, e.g., adaptive resolutions and small file sizes. In this paper, we propose MARVEL (MAnga’s Raster to VEctor Learning), a primitive-wise approach for vectorizing raster gray-level mangas by Deep Reinforcement Learning (DRL). Unlike previous learning-based methods which predict vector parameters for an entire image, MARVEL introduces a new perspective that regards an entire manga as a collection of basic primitives—stroke lines, and designs a DRL model to decompose the target image into a primitive sequence for achieving accurate vectorization. To improve vectorization accuracies and decrease file sizes, we further propose a stroke accuracy reward to predict accurate stroke lines, and a pruning mechanism to avoid generating erroneous and repeated strokes. Extensive subjective and objective experiments show that our MARVEL can generate impressive results and reaches the state-of-the-art level.

One of the key issues in mobile edge computing (MEC) is computation offloading, most policies of which are developed based on mathematical programming (MP). Due to the high computational complexity of iterative programming in MP-based policies, recent years have seen a popular trend to develop offloading policies based on deep reinforcement learning (DRL). However, on account of the poor generalization ability of DRL models in MEC environments with different network sizes and settings, it is difficult to directly apply DRL-based offloading policies in unseen MEC environments. Motivated by this, we propose a DRL-based environment-adaptive offloading framework (DEAT), including a size-adaptive scheme (SIED) and setting-adaptive component (SEAL). SIED leverages the idea of ‘time division multiplexing’ to adapt to varying MEC network sizes and order-unaware feature extraction to mitigate impacts of different size-changing orders. SEAL adopts system dynamics embedding and offloading policy embedding, which guide the finding of the closest pre-training MEC environment and offloading policy, respectively, to achieve fast setting-adaptation with only few exploring interactions in unseen MEC environments. Extensive experiments are conducted via both simulation and testbed to demonstrate the adaptation performance advantages of DEAT in unseen MEC environments compared to the state-of-the-art offloading approaches.

Cross-domain federated learning (FL), where data on local clients come from different domains, is a common case of FL. In such a cross-domain case, features extracted from the raw data of different clients deviate from each other in the feature space, leading to a so-called feature shift. This phenomenon can reduce feature discrimination and degrade the performance of the learned model. However, most existing FL methods are not specifically designed for the cross-domain setting. In this article, we propose a novel cross-domain FL method named AlignFed. In AlignFed, each client model consists of a personalized feature extractor and a shared lightweight classifier. The feature extractor maps the features to a consistent space by aligning them to identical global target points. Inspired by recent studies in contrastive learning, AlignFed regards points that are uniformly distributed on the hypersphere as global target points. It then pushes features toward global target points of their corresponding classes and away from those of other classes to improve feature discrimination. The shared classifier aggregates knowledge across clients over the consistent feature space, which can mitigate performance degradation caused by feature shift while reducing communication cost. We conduct convergence analysis and perform extensive experiments to evaluate AlignFed.

When doctors write a medical report, they first focus on key regions in the image, which usually contain abnormal information. Then, they write the report based on their experience and professional knowledge. Some existing medical report generation methods design corresponding schemes to model this process, but they still have some drawbacks. Some of them adopt the attention map to model the doctors’ image-reading behavior, which requires additional cropping and feature extraction operations, bringing many extra computations. Other methods manually construct knowledge graphs to model the doctors’ domain knowledge, which requires manual construction for different disease domains and has less generalization. Therefore, We propose DEKG, a doctors behavior aware and domain knowledge driven model, to improve the quality of generated reports by modeling the fixed pattern of doctors when writing reports. Specifically, feature clustering is utilized to model the doctors’ image-reading behavior that only brings fewer computations, and then we capture key regions with possible abnormalities. We also design a method to automatically construct domain knowledge graphs, which can quickly accomplish the construction process for different disease domains without human intervention. Extensive experiments on datasets from different disease domains demonstrate that DEKG achieves competitive results with state-of-the-art methods and can generate high-quality reports.

In the past decade, significant advancements have been made in utilizing deep learning for breast lesion segmentation. Recently, researchers have increasingly focused on harnessing the power of multiple modalities, recognizing its potential for enhancing segmentation performance. We observe that in clinical practice, many radiologists often rely on two types of ultrasound images, namely ultrasound (US) and contrast-enhanced ultrasound (CEUS) data for diagnosis. This motivates us to propose a multi-modal segmentation network, called as IMAN (Iterative Mutual-Aid Network), based on these two modalities. The architecture of IMAN adopts a novel hourglass shape, featuring two branches connected by an ‘X’ pathway. One branch is dedicated to processing CEUS data, while the other branch handles US data. Each branch generates segmentation results specific to its respective modality. The ’X’ pathway, realized by a margin mask generator module, serves as a bridge between these branches by forcing the segmentation results from one branch as additional input to the other. This head-to-tail pathway effectively facilitates mutual aid between the two modalities. In addition, we propose an iterative training policy during the training process to fully exploit the information from both US and CEUS data. Experimental results on a Breast-US-CEUS dataset comprising 169 samples demonstrate the effectiveness of IMAN, achieving Dice Similarity Coefficient of 83.96% and 81.16% for US images and CEUS videos, respectively. These scores surpass those obtained by many state-of-the-art segmentation methods. Furthermore, IMAN exhibits robust generalization capabilities across different segmentation structures.

Personalized federated learning (PFL) reduces the impact of non-independent and identically distributed (non-IID) data among clients by allowing each client to train a personalized model when collaborating with others. A key question in PFL is to decide which parameters of a client should be localized or shared with others. In current mainstream approaches, all layers that are sensitive to non-IID data (such as classifier layers) are generally personalized. The reasoning behind this approach is understandable, as localizing parameters that are easily influenced by non-IID data can prevent the potential negative effect of collaboration. However, we believe that this approach is too conservative for collaboration. For example, for a certain client, even if its parameters are easily influenced by non-IID data, it can still benefit by sharing these parameters with clients having similar data distribution. This observation emphasizes the importance of considering not only the sensitivity to non-IID data but also the similarity of data distribution when determining which parameters should be localized in PFL. This paper introduces a novel guideline for client collaboration in PFL. Unlike existing approaches that prohibit all collaboration of sensitive parameters, our guideline allows clients to share more parameters with others, leading to improved model performance. Additionally, we propose a new PFL method named FedCAC, which employs a quantitative metric to evaluate each parameter’s sensitivity to non-IID data and carefully selects collaborators based on this evaluation. Experimental results demonstrate that FedCAC enables clients to share more parameters with others, resulting in superior performance compared to state-of-the-art methods, particularly in scenarios where clients have diverse distributions. The code is integrated into our FL training framework: https://github.com/kxzxvbk/Fling.

Mobile edge computing (MEC) has been proposed as a promising paradigm to provide mobile devices with both satisfactory computing capacity and task latency. One key issue in MEC is computation offloading (CompOff), which has attracted numerous research interests. Most existing CompOff approaches are developed based on iterative programming (IterProg), that calculates a CompOff action based on system dynamics each time mobile tasks arrive. Due to the heavy dependency of IterProg on reliable system dynamics, as well as the online computational burden, recent years have seen a popular trend to develop CompOff approaches based on deep reinforcement learning (DRL), which could generate real-time model-free CompOff actions. However, due to the intrinsic poor generalization of DRL, it is hard to directly apply DRL-based policies in new MEC environments, and long-time fine-tuning is often required. To address the challenge, this paper proposes a fast transferable CompOff framework (named TransOff), based on the idea of embedded reinforcement learning. Specifically, TransOff is composed of multiple primitive CompOff policies (pCOPs) and a multiplicative composition function (MCF). The pCOPs and MCF are pre-trained in a diverse variety of MEC environments. When encountering new MEC environments, pCOPs are kept fixed to prevent catastrophic forgetting of pre-trained CompOff skills, while only MCF is fine-tuned to produce new compositions of pCOPs to achieve fast transfer. We conduct extensive experiments via both numerical simulation and real testbed, indicating the fast transfer ability of TransOff compared to the state-of-the-art DRL-based and meta learning-based CompOff approaches.

Multivariate time series anomaly detection, crucial for ensuring the safety of real-world systems, primarily focuses on extracting characteristics from time series under normal condition, and identifying potential anomalies throughout the evaluation process. Recent studies have achieved fruitful progress through mining the spatio-temporal relationships from multivariate time series, however, these approaches mostly neglect the latency among series which could lead to higher false alarm. Granger causality presents a promising solution to extract these inherent time-lagged relationships. Nonetheless, the intricate and dynamic relationships among numerous time series in real-world systems surpass the ability of linear Granger causality. To address this, we extend the linear Granger causality and propose the Granger Causal Former (GCFormer), a novel approach that leverages attention mechanisms to learn the inherent causal spatio-temporal relationships between historical and current timestamps across multiple time series. Specifically, GCFormer develops a Spatio-Mask (SM) to select the top-k most relevant series and a Temporal-Mask (TM) to concentrate attention on more recent historical timestamps. Moreover, to mitigate overfitting and ensure a smooth training process, GCFormer introduces an adjust top-k method and a TM penalty term. We evaluated GCFormer on four real-world benchmark datasets, demonstrating its superior performance over state-of-the-art approaches. Further analysis and a case study highlight the model’s novelty and interpretability.

As an effective approach for Vision Transformers (ViT) to obtain better initializations and representations in natural image analysis, Masked image modeling (MIM), performs the pretext task of reconstructing images by adopting partial observations without any label. Several works adopted dissimilar mask strategies to make ViT aggregate contextual information to infer missed contents. Nonetheless, chest radiographs conspicuously differ from photographic images, and conducting MIM in chest X-rays remains challenging. On that account, this paper came up with a specialized pre-training recipe cxrMIM and a masking strategy for chest radiographs on the basis of their physiological characters. In cxrMIM, the out-of-lung region was first analyzed, and the lung region was then reconstructed with the help of mechanical connections and similarities of anatomy and physiology. cxrMIM facilitates ViT to excavate the commonalities of pulmonary structures and promote better performance on downstream tasks. We conducted experiments on the ChestX-ray 14 dataset using advanced self-supervised methods (e.g. MoCo v3, MAE) for comparison. Quantitative and qualitative results signified that cxrMIM reinforced the efficiency of Vision Transformer to resolve the multi-label thorax disease classification problem, and cxrMIM pretrained Swin-B performed comparably to the state-of-the-art CNN models.

Dense video captioning, with the objective of describing a sequence of events in a video, has received much attention recently. As events in a video are highly correlated, leveraging relationships among events helps generate coherent captions. To utilize relationships among events, existing methods mainly enrich event representations with their context, either in the form of vision (i.e., video segments) or combining vision and language (i.e., captions). However, these methods do not explicitly exploit the correspondence between these two modalities. Moreover, the video-level context spanning multiple events is not fully exploited. In this paper, we propose MRCap, a novel relationship-based model for dense video captioning. The key of MRCap is a multi-modal and multi-level event relationship module (MMERM). MMERM exploits the correspondence between vision and language at both the event level and the video level via contrastive learning. Experiments on ActivityNet Captions and YouCook2 datasets demonstrate that MRCap achieves state-of-the-art performance.

Multi-hop question answering over incomplete knowledge graphs involves iteratively reasoning on the provided question and graph to find answers, while also tackling the inherent sparsity problem in the graph. Walk-based methods transform the reasoning process into a graph traversal task; however, they encounter challenges in convergence and stability due to the extensive action space and sensitivity to missing triples. On the other hand, embedding-based methods address the problem of missing triples but compromise interpretability in answer selection because of their black-box nature. We present GenBoost, a Generate-then-Boost framework for generative reasoning. By transforming the question-answering task into an inference path generation task, GenBoost effectively addresses existing limitations and offers a more efficient and interpretable approach for answer selection. GenBoost possesses two key features: (1) The reasoning procedure does not explicitly rely on the existing triples in the knowledge graph. By combining graph traversal and link prediction, our approach mitigates the impact of knowledge graph incompleteness. (2) Each entity in the reasoning path is generated autoregressively, providing insights into the decision-making process during multi-hop reasoning and enhancing interpretability. Extensive experiments conducted on incomplete knowledge graphs have demonstrated the effectiveness of our approach.

The recent success of extractive summarization depends on the availability of large-scale annotated datasets. Existing unsupervised approaches are mostly directed graph based by combining location information with centrality computing. These methods tend to generate summaries with two problems, one is low topic coverage of the source document called the facet bias problem, and the other is continuous position distribution of extracted sentences called the position bias problem. To solve these problems, we propose the topic-aware centrality-based sum-marization method (TACSUM). Specifically, we employ clustering techniques to explicitly model the topics of the document and define the metrics for topic consistency and topic coverage to improve the performance of summarization. The metric topic consistency is used to guide the calculation of centrality, which solves the position bias problem and achieves a more general effect in different scenarios. We combine the metric topic coverage with the centrality to enhance the topic awareness of the model, which ensures the selected sentences are important and diverse. Numerical experimental results on four datasets show that our method outperforms previous unsupervised methods, especially in long document domains. Extensive analyses confirm that our method can generate high-quality summaries by eliminating position bias and facet bias problems.

Federated learning (FL) trains a shared global model by periodically aggregating gradients from local devices. Communication overhead becomes a principal bottleneck in FL since participating devices usually suffer from limited bandwidth and unreliable connections in uplink transmission. To address this problem, the gradient compression methods based on compressed sensing (CS) theory have been put forward recently. However, most existing CS-based works compress gradients independently, ignoring the gradient correlations between participants or adjacent communication rounds, which constrains the achievement of higher compression rates. In view of the above observation, we propose a novel gradient compression scheme named FedDCS, guided by distributed compressed sensing (DCS) theory. Following the design philosophy of separate encoding and joint decoding in DCS, FedDCS compresses gradients for participants in each round separately while reconstructing them at the central server jointly via fully exploiting correlated gradients from the previous round, which are known as side information (SI). Benefiting from this design, reconstruction performance is significantly improved with fewer decoding errors also iterations under the identical compression rate, and the total uploading bits to achieve model convergence are considerably reduced. Theoretical analysis and extensive experiments conducted on MNIST and Fashion-MNIST both verify the effectiveness of our approach.

Federated learning (FL) is an emerging paradigm of distributed machine learning. However, when applied to wireless network scenarios, FL usually suffers from high communication cost because clients need to transmit their updated gradients to a server in every training round. Although many gradient compression techniques like sparsification and quantization are proposed, they compress clients’ gradients independently, without considering the correlations among gradients. In this paper, we propose SVDFed, a collaborative gradient compression framework for FL. SVDFed utilizes Singular Value Decomposition (SVD) to find a few basis vectors, whose linear combination can well represent clients’ gradients at a certain round. Due to the correlations among gradients, these basis vectors can still well approximate new gradients in many subsequent rounds. With the help of basis vectors, clients only need to upload the coefficients of the linear combination to the server, which greatly reduces communication cost. In addition, SVDFed leverages the classical PID (Proportional, Integral, Derivative) control to determine the proper time to update basis vectors to maintain their representation ability. Through experiments, we demonstrate that SVDFed outperforms existing gradient compression methods in FL. For example, compared to a popular gradient quantization method QSGD, SVDFed can reduce the communication overhead by 66 % and pending time by 99 %.

This paper investigates the use of reinforcement learning for autonomous exploration in an unknown environment. Autonomous exploration is crucial in many situations, such as urban search, security inspection, environmental mapping, etc. Traditional approaches focused on frontiers are unlikely to span a variety of enormously complex scenarios. Convergence is a little more difficult for learning-based approaches, which can adapt to many different environments. Consequently, a hierarchical exploration framework is built using frontier information. We propose a reinforcement learning-based local decision exploration model that uses deep neural networks to learn the optimal strategy from the environment. To prevent falling into local optimization, we also suggest a global rescue module to assist the robot in returning to the proper exploration track. Compared with other hierarchical methods, the framework is more effective and resilient in many contexts, greatly decreasing the total completion time and path length.

All six papers in this special section engage in different streams but extremely relevant domain vectors of Cognitive Cyber-Physical Systems (CCPS) with artificial intelligence (AI) based solutions in medical informatics. Highlights recent trends in the scientific community and presents emergent technologies, implementations, applications concerning the CPSS. CPSS is witnessing rapid transformation as an interdisciplinary technology that blends physical components and computing devices to enable AI-based solutions. CCPS will be playing a significant role that integrates machine learning/AI techniques and resulting in dramatic improvements for medical informatics and the future of human-augmentation. CPHMS coordinates supervisory medical systems and medical resources everywhere; there is a great scope towards health consciousness and healthy society. Medical Cyber-Physical Systems (MCPS) in healthcare towards critical integration in network of medical devices. MCPS is the next generation computing that is comprised of tightly coupled computational and communication components of medical automation systems such as clinical decision, early detection of health infectious, disease prevention, rapid analysis of health hazards and so on. CCPS and MCPS research would be created new models, new design, and integration models for large scale systems in comprehensive, holistic medical automation systems. With recent enlargements in the big data processing, cognitive data science and AI, it is now possible to create even more realistic digital twins that properly model different operating situations and characteristics to process the medical intelligence systems.

Screen lock is a critical security feature for smartphones to prevent unauthorized access. Although various screen unlocking technologies, including fingerprint and facial recognition, have been widely adopted, they still have some limitations. For example, fingerprints can be stolen by special material stickers and facial recognition systems can be cheated by 3D-printed head models. In this paper, we propose EmgAuth, a novel electromyography(EMG)-based smartphone unlocking system based on the Siamese network. EmgAuth enables users to unlock their smartphones by leveraging the EMG data of the smartphone users collected from Myo armbands. When training the Siamese network, we design a special data augmentation technique to make the system resilient to the rotation of the armband, which makes EmgAuth free of calibration. We conduct extensive experiments including 53 participants and the evaluation results verify that EmgAuth can effectively authenticate users with an average true acceptance rate of 91.81% while keeping the average false acceptance rate of 7.43%. In addition, we also demonstrate that EmgAuth can work well for smartphones with different screen sizes and for different scenarios. EmgAuth shows great promise to serve as a good supplement for existing screen unlocking systems to improve the safety of smartphones.

In order to better assist the rehabilitation treatment of patients with musculoskeletal injury, standard rehabilitation actions are needed to guide the musculoskeletal rehabilitation process. With more and more urgent demands, the musculoskeletal rehabilitation evaluation systems have attracted a high degree of attention. Experts have proposed a series of systems based on laser, ultrasound, and image, which can give reasonable recognition and judgment. However, these systems either require specialized and expensive equipment or can be affected by ionizing radiation. How to construct a musculoskeletal rehabilitation evaluation system with low cost, good effect, and little injury is still a great challenge. In this article, we propose MSEva, a musculoskeletal rehabilitation evaluation system based on EMG signals. Specifically, the system uses EMG sensors to collect a large amount of data for five rehabilitation actions. Secondly, MSEva uses Wavelet Transform (WT) to extract the signal features and then puts the processed data into the Long Short-Term Memory (LSTM) network for model training. Finally, the system uses the LSTM model to evaluate the normality of the EMG response of rehabilitation actions. The results show that the average accuracy of MSEva reaches 94.37%, which has important evaluation value in guiding the rehabilitation of musculoskeletal patients.

Automatically generating radiology reports has recently made great progress, which reduces the workload of radiologists. The high similarity among radiology images in training datasets forces existing methods to focus more on common medical facts existing in many images, such as pleural effusion and heart size. However, distinctive medical facts in individual images, such as edema and soft tissue that are rare in the whole dataset, receive little attention or are even ignored. This result leads to reports generated by existing methods that lack descriptions of distinctive medical facts in individual images. In this paper, we propose TRGD, a Transformer with graph-based distinctive attention, to generate high-quality radiology reports automatically. In TRGD, we first design a graph network to extract the features representing common medical facts. Then we lower their proportion in the global semantic features representing complete medical facts to increase the influence of distinctive medical facts. Experimental results on two datasets IU X-Ray and MIMIC-CXR demonstrate that, by sufficiently focusing on distinctive medical facts in individual images, TRGD indeed generates high-quality radiology reports and achieves competitive results with state-of-the-art methods.

Cephalometric tracing is widely used in orthodontic diagnosis and treatment planning. Since manual landmark lo-calization suffers from severe inter-observer and intra-observer inconsistency, a large number of efforts have been made by researchers to develop automatic localization methods. However, most of the existing methods are developed based on rules which sample uniformly from origin images rather than with the highest density in a focal point and ignore intermediate layers' results in their networks or their outputs' channels. To address the issue, this paper proposes a deep learning model based on multi-scale and multi-channel attention to identify landmarks. The channel attention network is first trained by multi-scale image patches cropped from 100 Cephalograms, and then enhanced by cross-layer connections to extract high-level features, finally involved in the collaboration with a three-layer MLP module to accurately locate coordinates. We conduct extensive evaluation on a real cephalometric X-ray data-set and a Non-public dataset, both achieve promising performance improvements especially in terms of high-precision detection.

Document-level Relation Extraction (DocRE) aims at extracting relations between entities in a given document. Since different mention pairs may express different relations or even no relation, it is crucial to identify key mention pairs responsible for the entity-level relation labels. However, most recent studies treat different mentions equally while predicting the relations between entities, leading to sub-optimal performance. To this end, we propose a novel DocRE model called Key Mention pairs Guided Relation Extractor (KMGRE) to directly model mention-level relations, containing two modules: a mention-level relation extractor and a key instance classifier. These two modules could be iteratively optimized with an EM-based algorithm to enhance each other. We also propose a new method to solve the multi-label problem in optimizing the mention-level relation extractor. Experimental results on two public DocRE datasets demonstrate that the proposed model is effective and outperforms previous state-of-the-art models.

Distantly supervised relation extraction aims to extract relational facts from texts but suffers from noisy instances. Existing methods usually select reliable sentences that rely on potential noisy labels, resulting in wrongly selecting many noisy training instances or underutilizing a large amount of valuable training data. This paper proposes a sentence-level DSRE method beyond typical instance selection approaches by preventing samples from falling into the wrong classification space on the feature space. Specifically, a theorem for denoising and the corresponding implementation, named Consensus Enhanced Training Approach (CETA), are proposed in this paper. By training the model with CETA, samples of different classes are separated, and samples of the same class are closely clustered in the feature space. Thus the model can easily establish the robust classification boundary to prevent noisy labels from biasing wrongly labeled samples into the wrong classification space. This process is achieved by enhancing the classification consensus between two discrepant classifiers and does not depend on any potential noisy labels, thus avoiding the above two limitations. Extensive experiments on widely-used benchmarks have demonstrated that CETA significantly outperforms the previous methods and achieves new state-of-the-art results.

Document-level relation extraction (DocRE)aims at extracting relations of all entity pairs in a document.A key challenge to DocRE lies in the complex interdependency between the relations of entity pairs.Unlike most prior efforts focusing on implicitly powerful representations, the recently proposed LogiRE (Ru et al., 2021) explicitly captures the interdependency by learning logical rules.However, Lo-giRE requires extra parameterized modules to reason merely after training backbones, and this disjointed optimization of backbones and extra modules may lead to sub-optimal results.In this paper, we propose MILR, a logic enhanced framework that boosts DocRE by Mining and Injecting Logical Rules.MILR first mines logical rules from annotations based on frequencies.Then in training, consistency regularization is leveraged as an auxiliary loss to penalize instances that violate mined rules.Finally, MILR infers from a global perspective based on integer programming.Compared with LogiRE, MILR does not introduce extra parameters and injects logical rules during both training and inference.Extensive experiments on two benchmarks demonstrate that MILR not only improves the relation extraction performance (1.1%-3.8%F1) but also makes predictions more logically consistent (over 4.5% Logic).More importantly, MILR also consistently outperforms LogiRE on both counts.Code is available at https:// github.com/XingYing-stack/MILR.

This Research to Practice Full Paper presented a Flipped Classroom (FC) approach to teaching a computer programming course. This approach increased students’ academic performance, course satisfaction and learning motivation by providing more time for active learning.There is an intense need for the studies of FC approach in higher education, especially in computer programming courses. Although FC pedagogy has achieved great success in K-12 schools, there is rare quantitative research on programming courses in higher education.This paper introduced a FC approach to teaching computer program course, and described in detail how to arrange pre-class activities and in-class activities. Pre-class activities were recommended to be arranged by time. 5 days before class, release course material and open access to programming practice. Students began to study and practice by themselves. 2 days before class, Students were required to complete pre-class test. 1 day before class, the teaching assistant provided a pre-class test analysis report. The lecturer prepared the coming class based on the report. In-class activities consisted of four parts. First, the lecturer explained in detail the common problems in the pre-class test. Second, the lecturer checked the problem-solving by asking questions. Third, students discussed in teams. Fourth, students gave presentations on the advanced topics.To assess the effectiveness of the FC approach, the study was conducted at the course Swift Language Programming Practice for undergraduate students. Students were divided into two groups. One group implemented traditional pedagogy, and the other group implemented FC pedagogy. The effect of the two groups was analyzed by teaching data collected during 2019-2021. The nonparametric independent-samples Kruskal–Wallis test was used to measure the changes in course performance and student satisfaction. The results showed that by introducing FC to teach computer programming courses, course performance (measured by examination results) and student satisfaction (measured by course questionnaires) were significantly improved.The contribution of this paper was to propose a FC approach suitable for college programming course, and through quantitative statistical methods to analyze the teaching effect of the course. The results showed that the approach significantly improved the course performance and student satisfaction. The FC approach of this paper can be regarded as a reference for similar course in college.

This Research to Practice Full Paper presented a comprehensive experiment approach to enhancing computer engineering ability. This approach integrated Swift programming language, iOS development, UML, software testing, MVC, Cocoa Touch Framework and Design Patterns into a comprehensive experiment, through which students can master the engineering methods to solve complex application problems.In college, traditional computer programming courses focus on the grammar and classical algorithm of programming language. Usually the amount of code is far lower than that of industrial products. Such programming courses can’t effectively improve students’ ability to solve complex engineering problems. They also can’t meet the requirements of industrial development. Students are not satisfied with the results of these courses. There is an intense need for the studies of enhancing student’s computer engineering ability.Taking Swift Language Programming course as an example, this paper presented a comprehensive experiment approach to enhancing students’ computer engineering ability by developing classic industrial iOS Apps.Flipped classroom pedagogy is conducive to free much time in class. Lecturers can fully communicate with students and help students complete challenging tasks. The comprehensive experiment consists of pre-class activities and in-class activities. Before class, the lecturer provides experiment materials online including theoretical handouts of Design Patterns, manuals of UML 2.0 specifications and Cocoa Touch reference manual, etc. Students learn the materials by themselves, practice and discuss online and complete the corresponding pre-class tests. In class, the lecturer analyzes in detail the problems students encounter after class and guides them to solve these problems. The lecturer also participates in each group discussion to ensure the smooth progress of students’ project.The implementation of comprehensive experiment is divided into four sub tasks. These tasks are app function analysis, App detailed design, programming implementation, and App release and launch. First, according to the requirements of the App, the function is analyzed in detail and defined with UML. Second, based on functional analysis, the App’s system architecture, data structure, view combination, logic execution process and core algorithms are designed. The system is defined in detail with UML Class diagram. Third, according to the detailed design of the App, user interface is built by Xcode storyboard, and the model layer, view layer and control layer are implemented in Swift. Then unit test and system test are conducted on the App and bugs are repaired. Finally, App launch is completed including App internationalization, developer certificate applying, creating description file, setting product identification and deployment information, and submitting App online.To assess the effect of this comprehensive experiment approach, three-year teaching data were analyzed using statistical methods. The results show that students’ engineering ability (measured by code scale) and student satisfaction (measured by questionnaires) were significantly improved.Our contribution is to propose a detailed comprehensive experiment approach to enhancing computer engineering ability. The analysis of teaching data show that it is helpful to improve students’ computer engineering ability and course satisfaction.

One vital challenge in federated learning (FL) is the statistical heterogeneity of data in different clients, which negatively affects the performance of the finally obtained model. One common approach to address this problem, called as personalized federated learning (PFL), is to train a personalized model for each client. A key design issue in PFL-based methods is determining which parts of the model should be personalized for each client. For example, one popular method in PFL is to personalize the batch normalization layers. In this paper, we propose ChannelFed, a new PFL-based method which personalizes the channel attention module. ChannelFed is designed based on the following observation: Channel attention assigns different weights to channels for different classes of data, which can be utilized to exploit knowledge of heterogeneous data from different clients. By keeping the channel attention module localized, ChannelFed enables clients to concentrate on client-specific channels. ChannelFed implements normalization across samples in the channel attention module to better fit for statistical heterogeneity scenarios. Experiments on CIFAR-10, Fashion-MNIST, and CIFAR-100 datasets demonstrate that ChannelFed outperforms other PFL methods under statistical heterogeneity scenarios.

Federated learning (FL) is an emerging machine learning paradigm where multiple distributed clients collaboratively train a model without centrally collecting their raw data. In FL setting, it is a common case that the data on local clients come from different domains, e.g., photos taken by different mobile phones can vary in intensity and contrast due to the difference of imaging parameters. In such a cross-domain case, features extracted from data of different clients deviate from each other in the feature space, leading to the so-called feature shift. The feature shift can reduce the discrimination of features and degrade the performance of the learned model. However, most existing FL methods are not particularly designed for cross-domain setting. In this paper, we propose a novel cross-domain FL method, named AlignFed. In AlignFed, the model on each client is separated to a personalized feature extractor and a shared classifier. The former extracts consistent features among clients by aligning features of different clients to some specific points in the feature space. The latter aggregates the knowledge across clients over the consistent feature space, which can mitigate the performance degradation caused by the feature shift in cross-domain FL. We conduct experiments on common-used multi-domain datasets, including Digits-Five, Office-Caltech10, and DomainNet. The experimental results demonstrate that AlignFed can outperform the state-of-art FL methods.

Recently, we made available WeNet [1], a production-oriented end-to-end speech recognition toolkit, which introduces a unified two-pass (U2) framework and a built-in runtime to address the streaming and non-streaming decoding modes in a single model.To further improve ASR performance and facilitate various production requirements, in this paper, we present WeNet 2.0 with four important updates.(1) We propose U2++, a unified two-pass framework with bidirectional attention decoders, which includes the future contextual information by a right-toleft attention decoder to improve the representative ability of the shared encoder and the performance during the rescoring stage.(2) We introduce an n-gram based language model and a WFSTbased decoder into WeNet 2.0, promoting the use of rich text data in production scenarios.(3) We design a unified contextual biasing framework, which leverages user-specific context (e.g., contact lists) to provide rapid adaptation ability for production and improves ASR accuracy in both with-LM and without-LM scenarios.(4) We design a unified IO to support large-scale data for effective model training.In summary, the brand-new WeNet 2.0 achieves up to 10% relative recognition performance improvement over the original WeNet on various corpora and makes available several important production-oriented features.

Data heterogeneity is one of the main challenges faced by federated learning (FL). Unlike traditional FL methods (e.g. FedAvg) which train a global model for all clients, personalized federated learning (PFL) can address the above problem by training a personalized model for each client. Current mainstream PFL researches first obtain a global model through collaborative training among all clients and then fine-tune the global model on each client's local data to obtain personalized models. However, this two-staged approach has a drawback: when the heterogeneity of different clients is large, the obtained final global model can deviate from the distributions of all clients, and therefore is not a good starting point for updating personalized models. In this paper, we propose pFedGF, a new PFL method based on gradient fusion. Different from traditional two-staged PFL, in each round of pFedGF, each client maintains two gradients simultaneously, a global gradient to capture information from all clients, and a local gradient that reflects the specific distribution of each client. The two gradients are fused to obtain the updated direction of the personalized model for each client. We carried out experiments on MNIST, FMNIST, and CIFAR-10 datasets. The results demonstrate that in the presence of data heterogeneity, pFedGF outperforms other PFL methods.

Infrastructure-assisted autonomous driving is an emerging paradigm that expects to significantly improve the driving safety of autonomous vehicles. The key enabling technology for this vision is to fuse LiDAR results from the roadside infrastructure and the vehicle to improve the vehicle's perception in real time. In this work, we propose VIPS, a novel lightweight system that can achieve decimeter-level and real-time (up to 100 ms) perception fusion between driving vehicles and roadside infrastructure. The key idea of VIPS is to exploit highly efficient matching of graph structures that encode objects' lean representations as well as their relationships, such as locations, semantics, sizes, and spatial distribution. Moreover, by leveraging the tracked motion trajectories, VIPS can maintain the spatial and temporal consistency of the scene, which effectively mitigates the impact of asynchronous data frames and unpredictable communication/compute delays. We implement VIPS end-to-end based on a campus smart lamppost testbed. To evaluate the performance of VIPS under diverse situations, we also collect two new multi-view point cloud datasets using the smart lamppost testbed and an autonomous driving simulator, respectively. Experiment results show that VIPS can extend the vehicle's perception range by 140% within 58 ms on average, and delivers a 4X improvement in perception fusion accuracy and 47X data transmission saving over existing approaches. A video demo of VIPS based on the lamppost dataset is available at https://youtu.be/zW4oi_EWOu0.

The century-old development of the automotive industry has spawned one of the greatest Cyber-Physical Systems (CPSs) in the future-unmanned vehicles. The vehicle can obtain environmental information through different sensors, map it to the virtual coordinate system of the vehicle body to make decisions, and finally generate control instructions. However, a series of factors, such as complex road scenes, defective and irregular target sparse sampling, and large coding space, pose challenges to accurate, efficient, and stable perception results. To overcome the most challenging problem of dynamic target tracking, this paper designs a two-stage detection model based on non-uniform polar voxelization sampling of irregular 3D point cloud, which is used with local registration-based search to achieve efficient multi-target tracking. Non-uniform voxelization not only balances the spatial sampling and encoding efficiency of the point cloud for the backbone, but also adapts to the feature aggregation of the detection head, thereby achieving double acceleration. Finally, we tested our model on KITTI Tracking data. The comparison results show that the calculation speed of the final model is greatly improved and the tracking accuracy is competitive in all categories.

Mobile devices (MDs) have undergone a booming development, yet are still capacity limited in computation and energy resources and thus could face troubles when serving computation-intensive and delay-sensitive applications. Mobile-edge computing (MEC) has been proposed to accommodate MDs with both satisfactory latency and acceptable resources, by offloading MDs’ tasks to near-deployed edge servers (ESs). Whereas, offloading tasks solely to ESs are difficult to meet distinct requirements of various applications, which leads to the emergence of end–edge–cloud orchestrated computing (EECOC). However, studies on EECOC are still insufficient, most existing of which do not consider MDs’ dynamical movements partly due to the intractability of associating optimal ESs with moving MDs. To address the issue, a novel deep reinforcement learning (DRL)-based mobility-aware (MA) EECOC scheduling approach is proposed in this article. With the goal of minimizing maximal task latency, we first formulate and transform the optimization problem into a Markov decision problem (MDP). Then, we enable DRL with elaborately designed reward functions and integrate it with NoisyNet to obtain near-optimal solutions. Furthermore, a MA component based on ConvLSTM is developed to extract MDs’ temporal–spatial distribution features and predict their movements, which are further utilized to facilitate the decision making of computation-offloading and resource-allocation actions. Extensive experimental results indicate the promising performance improvements of our approach against the state-of-the-art approaches in various scenarios.

To improve safety in public transportation, a major issue is how to avoid traffic accidents. To this end, a recent report has demonstrated that more than 90% of accidents in the United States were due to drivers’ abnormal behaviors. Relevant to this observation, many recent studies have proposed to use different sensors to monitor drivers’ behaviors and apply learning algorithms to detect abnormal behaviors. Nevertheless, most existing systems are expensive and inconvenient to be deployed or significantly affected by the environment. In this article, we propose and develop a novel and effective solution, namely, SafeDriving, that collects signals from electromyography (EMG) sensors and then utilizes an effective deep-learning model to detect abnormal behaviors in real time. Specifically, we first utilize a wearable EMG sensor that can be attached to a driver’s forearm to collect a large amount of sensing data from human drivers, for which we define five typical abnormal driving behaviors (i.e., fetching forward, picking up, turning the steering wheel sharply, turning back, and touching sunroof) and label each sample accordingly. Next, using the labeled data, we design and train multiple state-of-the-art classifiers to improve the performance of SafeDriving, e.g., convolutional neural network (CNN), long short-term memory (LSTM), and gated recurrent unit (GRU). The extensive experiments demonstrate that GRU can lead to the best performance with an average accuracy of 93.94%. Based on this observation, we further investigate other important factors, such as the binding area of the sensor, the tightness of binding, the duration of the sample, etc. The proposed SafeDriving system provides an effective approach to reliably assess drivers’ driving behaviors with affordable commodity sensors and be further used in public safety.

Every year, over 4700 traffic fatalities and 75000 crash injuries involve pedestrians in the United States. Effective solutions are urgently needed to prevent vehicle–pedestrian collision accidents. Many driving assistance systems are proposed to address this problem; however, they require additional infrastructures that may result in higher costs and be difficult to deploy on a large scale. In this article, we propose SafePath, which uses the ubiquitous smartphones to avoid vehicle–pedestrian collision. Specifically, SafePath utilizes the smartphones to broadcast the redesigned service set identifier (SSID) messages containing users’ information (e.g., location, direction, etc.) and scan the surroundings via wireless communications. Considering the limited communication range and the possible interference, and obstruction of obstacles, we propose a collaborative mechanism to enhance the transmission capability, hence predicting the collisions in advance effectively. We also design a risk evaluation scheme to calculate the probability of accidents and inform users to take actions against accidents at different levels. We implement SafePath on the Android platform and conduct extensive real-road experiments to evaluate the system performance. The experimental results demonstrate that SafePath can provide twice the transmission range compared with other collision-avoiding systems. Moreover, it also can significantly reduce the probability of vehicle–pedestrian collisions by up to 81.4%, with respect to other compared collision-avoiding systems in our real-road test.

Due to the high maneuverability and flexibility, unmanned aerial vehicles (UAVs) have been considered as a promising paradigm to assist mobile edge computing (MEC) in many scenarios including disaster rescue and field operation. Most existing research focuses on the study of trajectory and computation-offloading scheduling for UAV-assisted MEC in stationary environments, and could face challenges in dynamic environments where the locations of UAVs and mobile devices (MDs) vary significantly. Some latest research attempts to develop scheduling policies for dynamic environments by means of reinforcement learning (RL). However, as these need to explore in high-dimensional state and action space, they may fail to cover in large-scale networks where multiple UAVs serve numerous MDs. To address this challenge, we leverage the idea of “divide-and-conquer” and propose HT3O, a scalable scheduling approach for large-scale UAV-assisted MEC. First, HT3O is built with neural networks via deep RL to obtain real-time scheduling policies for MEC in dynamic environments. More importantly, to make HT3O more scalable, we decompose the scheduling problem into two-layered subproblems and optimize them alternately via hierarchical RL. This not only substantially reduces the complexity of each subproblem, but also improves the convergence efficiency. Experimental results show that HT3O can achieve promising performance improvements over state-of-the-art approaches.

Edge-supported Industrial Internet of Things (IIoT) has recently received significant attention since the edge computing can greatly improve the service quality of IIoT applications. However, edge servers are not fully trusted and are often deployed at the edge of the network. Therefore, there are some security challenges that need to be addressed. For edge-supported IIoT, a privacy-preserving range query is one of the most important functional requirements. Recently, some privacy-preserving range query solutions have been proposed in different fields. However, most of them only support single-dimensional range query, which are inefficient for the requirement of multidimensional range query. To address these problems, we propose a privacy-preserving multidimensional range query scheme for edge-supported IIoT, called Edge-PPMRQ, in this article. In Edge-PPMRQ, a novel range division algorithm is designed, through which the multidimensional ranges can be merged into one range, so as to achieve multidimensional range query through one query request. In addition, Edge-PPMRQ also supports the range queries for continuous, discontinuous, and arbitrary boundary ranges. The detailed security analysis proves that Edge-PPMRQ is privacy preserving for the query ranges, the query results, and the sensed data of IIoT devices. Furthermore, extensive comparison experiments also illustrate that Edge-PPMRQ is efficient in communication and computation.

The visual storytelling (VIST) task aims at generating reasonable, human-like and coherent stories with the image streams as input. Although many deep learning models have achieved promising results, most of them do not directly leverage the sentiment information of stories. In this paper, we propose a sentiment-aware generative model for VIST called SentiStory. The key of SentiStory is a multi-layered sentiment extraction module (MLSEM). For a given image stream, the higher layer gives coarse-grained but accurate sentiments, while the lower layer of the MLSEM extracts fine-grained but usually unreliable ones. The two layers are combined strategically to generate coherent and rich visual sentiment concepts for the VIST task. Results from both automatic and human evaluations demonstrate that with the help of the MLSEM, SentiStory achieves improvement in generating more coherent and human-like stories.

To realize flexible networking and on-demand topology reconstructing, software-defined industrial networks (SDINs) are increasingly embracing the flat structure. Similar to software defined networks (SDN), SDIN suffers from low traffic scheduling efficiency caused by large and imbalanced flows, known as the heavy hitters problem. Due to such heavy hitters, industrial networks may fail to satisfy application’s QoS requirements, which results in more severe damages. To improve flow scheduling efficiency under heavy hitters, this article introduces a novel imitation learning-based flow scheduling (ILFS) method. ILFS utilizes P4-based In-band Network Telemetry (INT) to collect fine-grained, real-time traffic data from SDIN’s data plane. In the control plane, it integrates the Generative Adversarial Imitation Learning (GAIL) model with a soft actor critic to preserve the experiences of flow, thereby better scheduling large flows. Our experiments thoroughly compare ILFS’s performance with several state-of-the-art traffic scheduling strategies. The results indicate that ILFS successfully controls the link bandwidth the utilization between 10$\%$and 80$\%$and significantly improves the average network throughput and link utilization rate.

In order to improve trajectory tracking accuracy, a reinforcement learning method was employed to address the trajectory tracking task in autonomous driving. There are many conveniences and advantages in theories, methods and tools utilizing reinforcement learning to solve trajectory tracking control problems. To teach an intelligent agent effective driving skills, the structured road information in the urban environment was extracted to generate an accurate reference trajectory. Then real-world scenarios were modeled to build a simulation environment for training. To effectively train the intelligent driving agent, Imitation Learning was firstly employed to teach the agent primary driving skills. Afterwards, Reinforcement Learning was adopted to optimize the agent’s driving policy. After the intelligent driving agent was well trained in the simulator, the tracking experiments were conducted in the simulator and the real-world scenarios. The proposed method was compared with base-line methods of geometric tracking, optimization-based tracking and learning-based tracking. The experimental results demonstrated that Reinforcement-Tracking can achieve accurate trajectory tracking performance and even exceed the accuracy of most baseline methods.

Object detection is the most critical and foundational sensing module for the autonomous movement platform. However, most of the existing deep learning solutions are based on GPU servers, which limits their actual deployment. We present an efficient multi-sensor fusion based object detection model that can be deployed on the off-the-shelf edge computing device for the vehicle platform. To achieve real-time target detection, the model eliminates a large number of invalid point clouds through ground filtering algorithm, and then adds texture information (fused from camera image) through point cloud coloring to enhance features. The proposed PV-EncoNet efficiently encodes both the spatial and texture features of each colored point through point-wise and voxel-wise encoding, and then predicts the position, heading and class of the objects. The final model can achieve about 17.92 and 24.25 Frame per Second (FPS) on two different edge computing platforms, and the detection accuracy is comparable with the state-of-the-art models on the KITTI public dataset (i.e., 88.54% for cars, 71.94% for pedestrians and 73.04% for cyclists). The robustness and generalization ability of the PV-EncoNet for the 3D colored point cloud detection task is also verified by deploying it on the local vehicle platform and testing it on real road conditions.

The low-rank matrix completion problem is fundamental in both machine learning and computer vision fields with many important applications, such as recommendation system, motion capture, face recognition, and image inpainting. In order to avoid solving the rank minimization problem which is NP-hard, several surrogate functions of the rank have been proposed in the literature. However, the matrix restored from the optimization problem based on the existing surrogate functions seriously deviates from the original one. In this paper, we first design a new non-convex Schatten capped$p$norm which generalizes several existing non-convex matrix norms and balances between the rank and the nuclear norm of the matrix. Then, a matrix completion method based on the Schatten capped$p$norm is proposed by exploiting the framework of the alternating direction method of multipliers. Meanwhile, the Schatten capped$p$norm regularized least squares subproblem is analyzed in detail and is solved explicitly. Finally, we evaluate the performance of the proposed matrix completion method based on extensive experiments in the field of image inpainting. All the experimental results demonstrate that the proposed method can indeed improve the accuracy of matrix completion compared with the existing methods.

As sharing of accounts is quite common among family members or roommates, the design of efficient recommender schemes for shared accounts has raised much attention recently. Generally speaking, after each login, it is essential for a recommender system to identify the current user behind and leverage this information to make recommendations. One naive approach is first to identify the identity of the current user and then make recommendations. However, this two-stage based approach may not achieve satisfactory performance. The key is that the recommended items favoring identifying users in the first stage may not be interesting to the users, which can deplete the user's patience quickly and cause early termination of users. To address the problem, we propose a novel recommendation scheme that makes a tradeoff between recommending discriminating items (helpful for identifying the user) and recommending interesting ones to the user (helpful for increasing the number of clicks). Under this scheme, we develop a patience model to capture the user's dynamic patience level during the recommendation process. Moreover, considering the increasing popularity of mobile devices, we also incorporate mobile sensor data (i.e., angle, accelerometer, gyroscope, etc.) into our approach to further improve the performance of the system. We implemented the above system in an App on mobile devices and carried out extensive experiments. The results demonstrate that our proposed scheme significantly outperforms the existing state-of-the-art approaches.

Mobile edge computing (MEC) has been envisioned as a promising paradigm that could effectively enhance the computational capacity of wireless user devices (WUDs) and quality of experience of mobile applications. One of the most crucial issues of MEC is computation offloading, which decides how to offload WUDs’ tasks to edge severs for further intensive computation. Conventional mathematical programming-based offloading approaches could face troubles in dynamic MEC environments due to the time-varying channel conditions (caused primarily by WUD mobility). To address the problem, reinforcement learning (RL) based offloading approaches have been proposed, which develop offloading policies by mapping MEC states to offloading actions. However, these approaches could fail to converge in large-scale MEC due to the exponentially-growing state and action spaces. In this article, we propose a novel online computation offloading approach that could effectively reduce task latency and energy consumption in dynamic MEC with large-scale WUDs. First, a RL-based computation offloading and energy transmission algorithm is proposed to accelerate the learning process. Then, a joint optimization method is adopted to develop the allocating algorithm, which obtains near-optimal solutions for energy and computation resources allocation. Simulation results show that the proposed approach can converge efficiently and achieve significant performance improvements over baseline approaches.

Convergecast is a fundamental operation in wireless sensor networks (WSNs). To support long-term deployment of WSNs, sensor nodes normally operate at low-duty-cycles. However, the low-duty-cycle operation significantly reduces the communication chance between nodes. Consequently, the risk of data collisions significantly increases when multiple senders transmit packets to a receiver during its very short active period. This problem further causes not only wasted packet retransmissions, but also a large delivery latency. Under such conditions, collision-free medium access is more appealing than recovering after collision for low-duty-cycle WSNs. In this work, we propose anincast-collision-free convergecast protocol, named iCore, to address the many-to-one collision problem in low-duty-cycle WSNs. iCore employs the dynamic forwarding technique, establishes a non-conflicting schedule for efficient convergecast, and improves the channel utilization by allowing senders to opportunistically transmit packets once detecting unused slots. Specifically, we design efficient forwarder assignment and forwarding optimization algorithms that ensure low end-to-end latency under diverse data traffic types. Through comprehensive performance evaluations, we demonstrate that, compared with the baseline protocol, iCore effectively minimizes the end-to-end delay by 25% ~ 57% and maintains high delivery ratio and energy efficiency for different many-to-one convergecast scenarios.

Manga is a world popular comic form originated in Japan, which typically employs black-and-white stroke lines and geometric exaggeration to describe humans' appearances, poses, and actions. In this paper, we propose MangaGAN, the first method based on Generative Adversarial Network (GAN) for unpaired photo-to-manga translation. Inspired by the drawing process of experienced manga artists, MangaGAN generates geometric features and converts each facial region into the manga domain with a tailored multi-GANs architecture. For training MangaGAN, we collect a new data-set from a popular manga work with extensive features. To produce high-quality manga faces, we propose a structural smoothing loss to smooth stroke-lines and avoid noisy pixels, and a similarity preserving module to improve the similarity between domains of photo and manga. Extensive experiments show that MangaGAN can produce high-quality manga faces preserving both the facial similarity and manga style, and outperforms other reference methods.

The performance of deep learning models generally relies on large and high-quality labeled datasets. However, in medical domain, as labeling process is much more laborious and time-consuming, most medical datasets are much smaller compared with natural image datasets. To mitigate this weakness, recent researches in medical image analysis adopt semi-supervised learning methods, especially consistency regularization methods to learn from a large amount of unlabeled medical data. However, as these semi-supervised learning methods are originally designed for tasks of natural images, specific properties of medical domain are not fully investigated and utilized. In this paper, we present DK-Consistency, a domain knowledge guided consistency regularization method for semi-supervised breast cancer diagnosis in ultrasound images. In DK-Consistency, domain knowledge of medical doctors is first incorporated into the generation process of perturbed samples for each unlabeled image. Then consistency regularization is adopted to force the model to make consistent predictions for unlabeled images and their perturbed samples. Extensive experiments demonstrate that, by injecting domain knowledge, DK-Consistency significantly improves the diagnostic performance of breast cancer and outperforms many state-of the-art semi-supervised methods.

In real application scenarios, the performance of deep networks may be degraded when the dataset contains noisy labels. Existing methods for learning with noisy labels are limited by two aspects. Firstly, methods based on the noise probability modeling can only be applied to class-level noisy labels. Secondly, others based on the memorization effect outperform in synthetic noise but get weak promotion in real-world noisy datasets. To solve these problems, this paper proposes a novel label-noise robust method named Discrepant Adversarial Training (DAT). The DAT method has ability of enforcing prominent feature extraction by matching feature distribution between clean and noisy data. Therefore, under the noise-free feature representation, the deep network can simply output the correct result. To better capture the divergence between the noisy and clean distribution, a new metric is designed to change the distribution divergence into computable. By minimizing the proposed metric with a min-max training of discrepancy on classifiers and generators, DAT can match noisy data to clean data in the feature space. To the best of our knowledge, DAT is the first to address the noisy label problem from the perspective of the feature distribution. Experiments on synthetic and real-world noisy datasets demonstrate that DAT can consistently outperform other state-of-the-art methods. Codes are available at https://github.com/Tyqnn0323/DAT.

Quick Response (QR) code is one of the most worldwide used two-dimensional codes. Traditional QR codes appear as random collections of black-and-white modules that lack visual semantics and aesthetic elements, which inspires the recent works to beautify the appearances of QR codes. However, these works adopt fixed generation algorithms and therefore can only generate QR codes with a pre-defined style. In this paper, combining the Neural Style Transfer technique, we propose a novel end-to-end method, named ArtCoder, to generate the stylized QR codes that are personalized, diverse, attractive, and scanning-robust. To guarantee that the generated stylized QR codes are still scanning-robust, we propose a Sampling-Simulation layer, a module-based code loss, and a competition mechanism. The experimental results show that our stylized QR codes have high-quality in both the visual effect and the scanning-robustness, and they are able to support the real-world application.

Relative position embedding (RPE) is a successful method to explicitly and efficaciously encode position information into Transformer models.In this paper, we investigate the potential problems in Shaw-RPE and XL-RPE, which are the most representative and prevalent RPEs, and propose two novel RPEs called Low-level Fine-grained High-level Coarse-grained (LFHC) RPE and Gaussian Cumulative Distribution Function (GCDF) RPE.LFHC-RPE is an improvement of Shaw-RPE, which enhances the perception ability at medium and long relative positions.GCDF-RPE utilizes the excellent properties of the Gaussian function to amend the prior encoding mechanism in XL-RPE.Experimental results on nine authoritative datasets demonstrate the effectiveness of our methods empirically.Furthermore, GCDF-RPE achieves the best overall performance among five different RPEs.

Factual inconsistencies existed in the output of abstractive summarization models with original documents are frequently presented.Fact consistency assessment requires the reasoning capability to find subtle clues to identify whether a model-generated summary is consistent with the original document.This paper proposes a fine-grained two-stage Fact Consistency assessment framework for Summarization models (SumFC).Given a document and a summary sentence, in the first stage, SumFC selects the top-K most relevant sentences with the summary sentence from the document.In the second stage, the model performs fine-grained consistency reasoning at the sentence level, and then aggregates all sentences' consistency scores to obtain the final assessment result.We get the training data pairs by data synthesis and adopt contrastive loss of data pairs to help the model identify subtle cues.Experiment results show that SumFC has made a significant improvement over the previous state-of-the-art methods.Our experiments also indicate that SumFC distinguishes detailed differences better.

Computer system ability training is a new trend in computer education. This paper proposed an innovative experimental teaching method of software and hardware collaborative design to better develop students' system view, structure view, engineering view of computers. An FPGA-based CNN accelerator was designed to combining the knowledge from software to compilation and then to hardware. The main innovations are: (1) Innovation of experimental system: A “curriculum tree” based on knowledge map was built to identify the implicit relationship between software and hardware knowledge. The curriculum was changed from Horizontal Teaching to Vertical Teaching in order to reduce the difficulties of cultivating system ability; (2) Innovation of experiment contents: it proposed an experiment teaching strategy of “Managing Complexity With Simplicity” guided by Occam's razor and used some effective methods to simply the system knowledge of each course around the top-level goals; (3) Innovation of experiment methods: a procedural and flow-based experiment model based on hierarchical experimental contents was used to achieve spiral progressive learning from software design to hardware simulation and then to system development; @Innovation of experiment platforms: an innovative method of conducting experiments, MODE (MODE = Experiment + MOOC) was proposed to allow students to do experiments “anytime, anywhere and on demand”.

Language modeling is an important problem in Natural Language Processing (NLP), and the multi-layer Transformer network is currently the most advanced and effective model for this task. However, there exist two inherent defects in its multi-head self-attention structure: (1) attention information loss: the lower-level attention weights cannot be explicitly passed through upper layers, which may lead the network lose some pivotal attention information captured by lower-level layers; (2) multi-head bottleneck: the dimension of each head in vanilla Transformer is relatively small and the process of each head is independent, which introduces an expressive bottleneck and makes subspace learning inadequate constitutionally. To overcome these two weaknesses, a novel neural architecture named Guide-Transformer is proposed in this paper. The Guide-Transformer utilizes horizontal and vertical attention information to guide the original process of the multi-head self-attention sublayer without introducing excessive complexity. The experimental results on three authoritative language modeling benchmarks demonstrate the effectiveness of Guide-Transformer. For the popular perplexity (ppl) and bits-per-character (bpc) evaluation metrics, Guide-Transformer achieves moderate improvements over the powerful baseline model.

The past year has witnessed the rapid development of applying the Transformer module to vision problems. While some researchers have demonstrated that Transformer-based models enjoy a favorable ability of fitting data, there are still growing number of evidences showing that these models suffer over-fitting especially when the training data is limited. This paper offers an empirical study by performing step-by-step operations to gradually transit a Transformer-based model to a convolution-based model. The results we obtain during the transition process deliver useful messages for improving visual recognition. Based on these observations, we propose a new architecture named Visformer, which is abbreviated from the ‘Vision-friendly Transformer’. With the same computational complexity, Visformer outperforms both the Transformer-based and convolution-based models in terms of ImageNet classification accuracy, and the advantage becomes more significant when the model complexity is lower or the training set is smaller. The code is available at https://github.com/danczs/Visformer.

LiDAR-only object detection is essential for autonomous driving systems and is a challenging problem. For the representation of a bird’s eye view LiDAR point-cloud, this paper proposes a single-stage object detector. The detector can output classification information and accurate positioning information for multi-category objects. In this paper, the detector’s design methods are detailed from a bird’s eye view LiDAR point-cloud encoding, network design, data augmentation, etc. The detector was evaluated on three challenging datasets: KITTI, nuScenes and Waymo. The experimental results demonstrated that the proposed detector can accurately achieve object detection tasks and the detection speed can reach 26.9 FPS. Both the precision and the speed can meet the requirements of most autonomous driving scenarios.

Quick Response (QR) code is a popular form of matrix barcodes that are widely used to tag online links on print media (e.g., posters, leaflets, and books). However, standard QR codes typically appear as noise-like black/white squares (named modules) which seriously disrupt the attractiveness of their carriers. In this paper, we propose StyleCode-Net, a method to generate novel art-style QR codes which can better match the entire style of their carriers to improve the visual quality. For endowing QR codes with artistic elements, a big challenge is that the scanning-robustness must be preserved after transforming colors and textures. To address these issues, we propose a module-based deformable convolutional mechanism (MDCM) and a dynamic target mechanism (DTM) in StyleCode-Net. MDCM can extract the features of black and white modules of QR codes respectively. Then, the extracted features are fed to DTM to balance the scanning-robustness and the style representation. Extensive subjective and objective experiments show that our art-style QR codes have reached the state-of-the-art level in both visual quality and scanning-robustness, and these codes have the potential to replace standard QR codes in real-world applications.

Recent years have witnessed the increasing popularity of mobile applications, e.g., virtual reality, unmanned driving, which are generally computation-intensive and latency-sensitive, posing a major challenge for resource-limited user equipment (UE). Mobile edge computing (MEC) has been proposed as a promising approach to alleviate the problem, by offloading mobile tasks to the edge server (ES) deployed in close proximity to UE. However, most existing task offloading algorithms are primarily based on centralized scheduling, which could suffer from the ‘curse of dimensionality’ in large MEC environments. To address this issue, this paper proposes a fully distributed task offloading approach based on multi-agent deep reinforcement learning, whose critic and actor neural networks are trained under the assistance of global and local network states, respectively. In addition, we design a model parameter aggregation mechanism, along with a normalized fine-tuned reward function, to further improve the learning efficiency of the training process. Simulation results show that our proposed approach could achieve substantial performance improvements over baseline approaches.

Large-scale canals with cascaded pools are constructed wordwide to divert water from rich to arid areas to mitigate water shortages. Efficient control of canals is essential to improve water-diversion performance. Numerous model-based approaches have been proposed and made great progress for canal control. However, when the predictive model is unavailable or unpromising for long time step predictions, model-free approaches could be considered as a possible way to achieve efficient control. Since most existing model-free approaches are focused on control of small canals or reservoirs, this article proposes a new control approach named policy and action reinforcement learning (PARL) for large-scale canals. We leverage the idea of “divide and conquer” to decompose the control task of large-scale canals into policy learning and action learning subtasks, and develop PARL by means of hierarchical reinforcement learning. Extensive experiments are conducted via numerical simulation on the case study of Chinese South to North Water Transfer Project, and experimental results show that PARL can achieve desirable performance improvements over other model-free learning approaches.

Autonomous vehicles need to solve the road keeping problem and the existing solutions based on reinforcement learning are mainly implemented in the simulators. The key of transferring the well-trained models to the real world is bridging the gaps between the simulator scenarios and the real scenarios. In this paper, we propose a method called ReinforcementDriving which explores navigation skills and trajectories from simulator for full-sized road keeping. Based on the real scenario, a driving simulator is firstly established to train an intelligent driving agent. The well-trained ReinforcementDriving agent is evaluated in a real-world scenario. We compare our work with human driving, optimal control-based tracking methods and other reinforcement learning-based lane following methods. The results demonstrate that the ReinforcementDriving system can effectively achieve lane keeping in a realistic scenario with satisfactory running time and lateral accuracy.

Online banks may disclose consumers’ shopping preferences due to various attacks. With differential privacy, each consumer can disturb his consumption amount locally before sending it to online banks. However, directly applying differential privacy in online banks will incur problems in reality because existing differential privacy schemes do not consider handling the noise boundary problem. In this paper, we propose an Optimized Differential prIvate Online tRansaction scheme (O-DIOR) for online banks to set boundaries of consumption amounts with added noises. We then revise O-DIOR to design a RO-DIOR scheme to select different boundaries while satisfying the differential privacy definition. Moreover, we provide in-depth theoretical analysis to prove that our schemes are capable to satisfy the differential privacy constraint. Finally, to evaluate the effectiveness, we have implemented our schemes in mobile payment experiments. Experimental results illustrate that the relevance between the consumption amount and online bank amount is reduced significantly, and the privacy losses are less than 0.5 in terms of mutual information.

In recent years, deep learning has been widely used in breast cancer diagnosis, and many high-performance models have emerged. However, most of the existing deep learning models are mainly based on static breast ultrasound (US) images. In actual diagnostic process, contrast-enhanced ultrasound (CEUS) is a commonly used technique by radiologists. Compared with static breast US images, CEUS videos can provide more detailed blood supply information of tumors, and therefore can help radiologists make a more accurate diagnosis. In this paper, we propose a novel diagnosis model based on CEUS videos. The backbone of the model is a 3D convolutional neural network. More specifically, we notice that radiologists generally follow two specific patterns when browsing CEUS videos. One pattern is that they focus on specific time slots, and the other is that they pay attention to the differences between the CEUS frames and the corresponding US images. To incorporate these two patterns into our deep learning model, we design a domain-knowledge-guided temporal attention module and a channel attention module. We validate our model on our Breast-CEUS dataset composed of 221 cases. The result shows that our model can achieve a sensitivity of 97.2% and an accuracy of 86.3%. In particular, the incorporation of domain knowledge leads to a 3.5% improvement in sensitivity and a 6.0% improvement in specificity. Finally, we also prove the validity of two domain knowledge modules in the 3D convolutional neural network (C3D) and the 3D ResNet (R3D).

Quick response (QR) codes are usually scanned in different environments, so they must be robust to variations in illumination, scale, coverage, and camera angles. Aesthetic QR codes improve the visual quality, but subtle changes in their appearance may cause scanning failure. In this article, a new method to generate scanning-robust aesthetic QR codes is proposed, which is based on a module-based scanning probability estimation model that can effectively balance the tradeoff between visual quality and scanning robustness. Our method locally adjusts the luminance of each module by estimating the probability of successful sampling. The approach adopts the hierarchical, coarse-to-fine strategy to enhance the visual quality of aesthetic QR codes, which sequentially generate the following three codes: a binary aesthetic QR code, a grayscale aesthetic QR code, and the final color aesthetic QR code. Our approach also can be used to create QR codes with different visual styles by adjusting some initialization parameters. User surveys and decoding experiments were adopted for evaluating our method compared with state-of-the-art algorithms, which indicates that the proposed approach has excellent performance in terms of both visual quality and scanning robustness.

Numerous research efforts have been conducted to simulate the crowd movements, while relatively few of them are specifically focused on multihazard situations. In this paper, we propose a novel crowd simulation method by modeling the generation and contagion of panic emotion under multihazard circumstances. In order to depict the effect from hazards and other agents to crowd movement, we first classify hazards into different types (transient and persistent, concurrent and nonconcurrent, and static and dynamic) based on their inherent characteristics. Second, we introduce the concept of perilous field for each hazard and further transform the critical level of the field to its invoked-panic emotion. After that, we propose an emotional contagion model to simulate the evolving process of panic emotion caused by multiple hazards. Finally, we introduce an emotional reciprocal velocity obstacles (RVOs) model to simulate the crowd behaviors by augmenting the traditional RVO model with emotional contagion, which for the first time combines the emotional impact and local avoidance together. Our experimental results demonstrate that the overall approach is robust, can better generate realistic crowds and the panic emotion dynamics in a crowd. Furthermore, it is recommended that our method can be applied to various complex multihazard environments.

Automatic designing computationally efficient neural networks has received much attention in recent years. Existing approaches either utilize network pruning or leverage the network architecture search methods. This paper presents a new framework named network adjustment, which considers network accuracy as a function of FLOPs, so that under each network configuration, one can estimate the FLOPs utilization ratio (FUR) for each layer and use it to determine whether to increase or decrease the number of channels on the layer. Note that FUR, like the gradient of a non-linear function, is accurate only in a small neighborhood of the current network. Hence, we design an iterative mechanism so that the initial network undergoes a number of steps, each of which has a small 'adjusting rate' to control the changes to the network. The computational overhead of the entire search process is reasonable, i.e., comparable to that of re-training the final model from scratch. Experiments on standard image classification datasets and a wide range of base networks demonstrate the effectiveness of our approach, which consistently outperforms the pruning counterpart. The code is available at https://github.com/danczs/NetworkAdjustment.

The existing deep CNN models on single image super-resolution processing are computationally-intensive in terms of memory usage and training time. In resources-limited platforms, it is desirable to consider developing light-weight models for super-resolution tasks. This paper proposes a parallel-group convolution, which uses 25% computation of the standard convolutions. With parallel-group convolutions, we develop an efficient light-weight convolutional neural network named MobileSR for super-resolution. Experimental results show that our proposed method achieves appreciable improvements over the state-of-the-art models with approximately 75% size reduction. The source code is available at https://github.com/DestinyK/MobileSR.

Online bank system has been widely deployed to provide financial service worldwide. Big data mining brings serious privacy issues to consumers because consumption records may disclose their activities. Existing authentication and encryption algorithms in privacy-preserving deposit transaction schemes are mostly limited to the case that each consumption has to be reported to online banks directly and an insider adversary can monitor a consumer's bank account during the financial transaction period. Motivated by the above problems, we propose a deposit transaction scheme based on the differential privacy. Considering a mobile payment system as a noise generator, this scheme aims to reduce the relevancy between real consumption and deposit transaction. In addition, we provide in-depth theoretical analysis that our scheme can satisfy the definition of the differential privacy. Experimental results illustrate the consumption pattern can be protected with the differential privacy.

Convolutional neural networks (CNNs) have achieved remarkable success in image recognition. Although the internal patterns of the input images are effectively learned by the CNNs, these patterns only constitute a small proportion of useful patterns contained in the input images. This can be attributed to the fact that the CNNs will stop learning if the learned patterns are enough to make a correct classification. Network regularization methods like dropout and SpatialDropout can ease this problem. During training, they randomly drop the features. These dropout methods, in essence, change the patterns learned by the networks, and in turn, forces the networks to learn other patterns to make the correct classification. However, the above methods have an important drawback. Randomly dropping features is generally inefficient and can introduce unnecessary noise. To tackle this problem, we propose SelectScale. Instead of randomly dropping units, SelectScale selects the important features in networks and adjusts them during training. Using SelectScale, we improve the performance of CNNs on CIFAR and ImageNet.

Lidar and optical camera are common sensors in the sensor layer of autopilot system. Lidar can use depth data to obtain accurate relative distance and contour information of obstacles, which is not easily affected by external light conditions. Optical camera can obtain rich object/environment semantic information through high-resolution image, which is relatively mature in technology. The two different sensors are highly complementary, and previous studies show that the fusion of laser point cloud and image data can greatly improve the efficiency of object detection in out door environment. In this paper, a deep convolutional neural network detection model based on Lidar and image information features layered fusion is studied. We try different fusion depth at the CNN model to seek the best solution according to the detection performance. The experimental results on the KITTI dataset show that the detection accuracy of the fusion based on YOLOv3 is 1.08% higher than original model. Another small scale experiment with our own self-driving platform on local area also show the final fusion model can achieve better detection accuracy in real road condition.

Pyroelectric infrared (PIR) sensors are considered to be promising devices for device-free localization due to its advantages of low cost, energy efficiency, and the immunity from multi-path fading. However, most of the existing PIR-based localization systems only utilize the binary information of PIR sensors and therefore require a large number of carefully deployed PIR sensors. A few works directly map the raw data of PIR sensors to one's location using machine learning approaches. However, these data-driven approaches require abundant training data and suffer from environmental change. In this paper, we propose PIRATES, a PIR-based device-free localization system based on the raw data of PIR sensors. The key of PIRATES is to extract a new type of location information called azimuth change. The extraction of the azimuth change relies on the physical properties of PIR sensors. Therefore, no abundant training data are needed and the system is robust to environmental change. Through experiments, we demonstrate that PIRATES can achieve higher localization accuracy than the state-of-the-art approaches. In addition, the information of the azimuth change can be easily incorporated with other information of PIR signals (e.g. amplitude) to improve the localization accuracy.

Screen lock is a critical security feature for smart-phones to prevent unauthorized access. Although various screen unlocking technologies including fingerprint and facial recognition have been widely adopted, they still have some limitations. For example, fingerprints can be stolen by special material stickers and facial recognition systems can be cheated by 3D-printed head models. In this paper, we propose EmgAuth, a novel electromyography(EMG)-based smartphone unlocking system based on the Siamese network. EmgAuth leverages the Myo armband to collect the EMG data of smartphone users and enables users to unlock their smartphones when picking up and watching their smartphones. In particular, when training the Siamese network, we design a special data augmentation technique to make the system resilient to the rotation of the armband. We conduct experiments including 40 participants and the evaluation results show that EmgAuth can effectively authenticate users with an average true acceptance rate of 91.81% while keeping the average false acceptance rate of 7.43%. In addition, we also demonstrate that EmgAuth can work well for smartphones with different sizes and at different locations, and is applicable for users with different postures. EmgAuth bears great promise to serve as a good supplement for existing screen unlocking systems to improve the safety of smartphones.

Due to easy access to smartphones, recent years have witnessed an increasing interest in using the mobile phone as a sensing and computation platform for vehicle steering detection. However, relatively lower accuracy of smartphone sensors than on-board diagnostic (OBD)-based systems often leads to lower accuracy. We propose an ensemble learning-based model combined with the heuristic algorithm for smartphone-based vehicle steering detection in this paper. Ensemble learning has been widely recognized for its powerful generalization capability, high accuracy, and rapid convergence. However, applying the ensemble learning approach to steering detection of the smartphone-based vehicle entails many challenges due to the limitation of smartphone storage, the constraint on power consumption, and the requirement of being real-time. To address these challenges, we propose a series of techniques to reduce the complexity of the model and energy consumption, while at the same time maintaining high detection accuracy. The performance of the proposed system has been demonstrated using a real dataset and can achieve an accuracy of 97.37%. We also conduct two case studies on real road environment in Beijing with different smartphones.

Twitter sentiment analysis has become a hot research topic in recent years. Most of existing solutions to Twitter sentiment analysis basically only consider textual information of Twitter messages, and struggle to perform well when facing short and ambiguous Twitter messages. Recent studies show that sentiment diffusion patterns on Twitter have close relationships with sentiment polarities of Twitter messages. Therefore, in this paper, we focus on how to fuse textual information of Twitter messages and sentiment diffusion patterns to obtain better performance of sentiment analysis on Twitter data. To this end, we first analyze sentiment diffusion by investigating a phenomenon called sentiment reversal, and find some interesting properties of sentiment reversals. Then, we consider the inter-relationships between textual information of Twitter messages and sentiment diffusion patterns, and propose an iterative algorithm called SentiDiff to predict sentiment polarities expressed in Twitter messages. To the best of our knowledge, this work is the first to utilize sentiment diffusion patterns to help improve Twitter sentiment analysis. Extensive experiments on real-world dataset demonstrate that compared with state-of-the-art textual information based sentiment analysis algorithms, our proposed algorithm yields PR-AUC improvements between 5.09 and 8.38 percent on Twitter sentiment classification tasks.

Due to the unavailability of Vehicle-to-Infrastructure (V2I) communication in current transportation systems, Traffic Light Detection (TLD) is still considered an important module in autonomous vehicles and Driver Assistance Systems (DAS). To overcome low flexibility and accuracy of vision-based heuristic algorithms and high power consumption of deep learning-based methods, we propose a lightweight and real-time traffic light detector for the autonomous vehicle platform. Our model consists of a heuristic candidate region selection module to identify all possible traffic lights, and a lightweight Convolution Neural Network (CNN) classifier to classify the results obtained. Offline simulations on the GPU server with the collected dataset and several public datasets show that our model achieves higher average accuracy and less time consumption. By integrating our detector module on NVidia Jetson TX1/TX2, we conduct on-road tests on two full-scale self-driving vehicle platforms (a car and a bus) in normal traffic conditions. Our model can achieve an average detection accuracy of 99.3 percent (mRttld) and 99.7 percent (Rttld) at 10Hz on TX1 and TX2, respectively. The on-road tests also show that our traffic light detection module can achieve <; + 1:5m errors at stop lines when working with other selfdriving modules.

Despite many existing research on data forwarding in low-duty-cycle wireless sensor networks (WSNs), relatively little work has been done on energy-efficient data forwarding with probabilistic delay bounds. Probabilistic delay guarantees (i.e., delay bounded data delivery with reliability constraints) are of increasing importance for many delay-constrained applications, since deterministic delay bounds are prohibitively expensive to guarantee in WSNs. However, radio duty-cycling and unreliable wireless links pose challenges for achieving the probabilistic delay guarantee in WSNs. In this paper, we propose EEAF, a novel energy-efficient adaptive forwarding technique tailored for low-duty-cycle WSNs with unreliable wireless links. We show the existence of path diversity in low-duty-cycle WSNs, where delay-optimal routing and energy-optimal routing are likely following different paths. The key idea of EEAF is to exploit the intrinsic path diversity to provide probabilistic delay guarantees while minimizing transmission cost. In EEAF, an early arriving packet will be adaptively switched to the energy-optimal path for energy conservation. Delay quantiles are derived at each node in a distributed manner and are used as the guidelines in the adaptive forwarding decision making. Extensive testbed experiment and large-scale simulation show that EEAF effectively reduces the transmission cost by 12%~25% with probabilistic delay guarantees under various network settings. In addition, we extend the EEAF technique with data aggregation for event-based traffic scenarios. Evaluation using publicly available WSN event traffic traces yields very encouraging results with up to 40% energy saving in probabilistic delay bounded data delivery.

The fourth industrial revolution (Industry 4.0) has promoted the all-around transformation of education in engineering. To cope with the changes caused by Industry 4.0, the Ministry of Education of China announced a new strategic guideline named “Emerging Engineering Education”, which proposed new demands of C programming courses. It is required to pay more attention to cultivate students' engineering practice ability. However, currently there is a phenomenon of “disconnection between teaching and application” behaving as some educators focus on teaching knowledge, but ignore training students' engineering skills, engineering thinking and engineering accomplishment. In order to solve these problems, we establish a new engineering-oriented curriculum system to cultivate interdisciplinary talents with strong engineering practice ability by designing innovative teaching pattern, advanced teaching methods and optimized teaching goals. The main measures are: 1) To construct a three-in-one curriculum system of “experiment, practice and internship”. It aims to implement engineering-oriented teaching strategies by designing course contents guided by enterprise needs and taking project as the carrier. 2) To construct an engineering-oriented teaching pattern of “multiple subjects, double tutors”. It wants to establish a netlike teaching model including teaching community of “teacher-student”, learning community of “student-student”, practice community of “student-enterprise” and guiding community of “intramural advisor-extramural advisor”. 3) To construct project-driven teaching contents. It commits to design a teaching chain of “knowledge + experiment + project” following the learning rule of “from perceptual-knowledge to rational-knowledge to practice-knowledge”. In conclusion, this paper proposed a new engineering application-oriented curriculum system, which takes engineering projects as cases, takes the cycle of project development as main-line and takes the cultivation of engineering talents as goals.

Target-level sentiment classification aims to identify the sentiment of multiple targets in a sentence. Although existing approaches based on neural network have achieved good performance in this task, we find that many approaches tend to give the same predictions for instances that have multiple targets, and this tendency can make a low accuracy for those instances that have different classes for different targets. Based on this observation, we propose MemNetAR, a memory network which can explicitly leverage the adversative relation among multiple targets in a sentence. Specifically, we add an adversative loss to the cross-entropy loss when there are adversative words between targets. The experimental results on public laptop and restaurant datasets prove that our model can improve 0.84% and 0.8% on total test dataset, and improve 2.97% and 2.68% on the dataset consisting of those instances with multiple targets but different classes by leveraging this new adversative information.

Text clustering has been widely used in many Natural Language Processing (NLP) applications such as text summarization and news recommendation. However, most of the current algorithms need to predefine a clustering number, which is difficult to obtain. Moreover, the mutli-label clustering is useful in multiple clustering tasks in many applications, but related works are rarely available. Although several studies have attempted to solve above two problems, there is a need for methods that can solve the two issues simultaneously. Therefore, we propose a new text clustering algorithm called Word2Cluster. Word2Cluster can automatically generate an adaptive number of clusters and support multi-label clustering. To test the performance of Wrod2Cluster, we build a Chinese text dataset, Hotline, according to real world applications. To evaluate the clustering results better, we propose an improved evaluation method based on basic accuracy, precision and recall for multi-label text clustering. Experimental results on a Chinese text dataset (Hotline) and a public English text dataset (Reuters) demonstrate that our algorithm can achieve better F1-measure and runs faster than the state-of- the-art baselines.

Gesture recognition systems based on myoelectric signal have raised more and more attention from researchers. Traditional gesture recognition methods are susceptible to multiple types of noise and require a large number of features, which increase overhead and decrease recognition efficiency. Fully utilizing the characteristics of the signal to recognize gestures is a big challenge. This paper proposes an improved empirical mode decomposition method (XB-EMD) based on autocorrelation function to denoise myoelectric signal. In addition, a novel deep neural network (CRNet) which combines the CNN and RNN layers together is trained for classifying the gestures based on denoised myoelectric signal. Experimental results show that the proposed gesture recognition system can improve recognition effectiveness at 97.4% with 10 typical gestures, and 72.55% with 16 complicated gestures.

Recently, the automatic text summarization has been widely used in text compression tasks. The Attention mechanism is one of the most popular methods used in the seq2seq (Sequence to Sequence) text summarization models. The current attention mechanisms usually use the hidden states of the encoder and the decoder to generate attention distributions. However, they ignore the information of the word waiting to be input into the decoder, leading to possible failures to obtain accurate attention distributions. In this work, we propose a novel attention mechanism further adding the decoder inputs into the operation of generating attention distributions. To our best knowledge, this is the first time that the decoder input has been added to the process of calculating the attention vector. The attention mechanism we proposed to generate the attention distributions considers context similarities as well as semantic similarities, which is closer to the behavior of the human summarizer. We also applied our attention mechanism to the seq2seq based summarization model and trained it on a large corpus containing hundreds of thousands of article-summary pairs. The experimental results on two summarization datasets demonstrate that our attention mechanism outperforms the existing well-known ones. For the popular evaluation metric of the text summarization, our method obtains a 2.93 ROUGE-2 score relative gain compared with the popular attention mechanism Bahdanau Attention, and a 2.21 ROUGE-2 score improvement compared with the best baseline method Luong Attention.

With the blossoming of social networking platforms like Twitter and Facebook, how to infer the opinions of online social network users on specific topics they had not directly given yet, has received much attention. Existing solutions mainly rely on one's previous posted messages. However, recent studies show that over 40% of users opt to be silent all or most of the time and post very few messages. Consequently, the performance of existing solutions will drop dramatically when they are applied to infer silent users' opinions, and how to infer the opinions of these silent users becomes a meaningful while challenging task. Inspired by the collaborative filtering techniques in cold-start recommendations, we infer the opinions of silent users by leveraging the text content posted by active users and their relationships between silent users. Specifically, we first consider both observed and pseudo relationships among users, and cluster users into communities in order to extract various kinds of features for opinion inference. We then design a coupled sparse matrix factorization (CSMF) model to capture the complex relations among these features. Extensive experiments on real-world data from Twitter show that our CSMF model achieves over 80% accuracy for the inference of silent users' opinions.

Internet of Things (IoT) can be used to promote many advanced applications by utilizing the sensed data collected from various settings. To reduce the energy consumption of IoT devices, and to extend the lifetime of network, the sensed data are usually compressed before their transmission through compressed sensing theory. By reconstructing the sensed data at the edge of network with more resourceful devices, such as laptops and servers, the intensive computation and energy consumption of the IoT nodes could be effectively offloaded. However, most of the existing data collection schemes are limited in their scalability, because the unified data reconstruction models of them are not suitable for large-scale surveillance scenarios. In our proposed scheme, the whole network is first partitioned into a number of data correlated clusters based on spatial correlation. Then, a data collection tree is built to collect the compressed data in a hybrid mode. Finally, the data reconstruction problem is modelled as a group sparse problem and solved through using an alternating direction method of multiplier-based algorithm. The performance of data communication and reconstruction of the proposed scheme is evaluated through experiments with real data set. The experimental results show that the proposed scheme can indeed lower the amount of data transmission, prolong the network life, and achieve a higher level of accuracy in data collection compared to existing data collection schemes.

Music mood is useful for music-related applications such as music retrieval or recommendation, which represents the inherent emotional expression of music signals. In this paper, a novel technique is proposed for music signal analysis in the view of emotions, which is based on the orthogonal matching pursuit, Gabor functions, and the Wigner distribution function. The technique, called the OMPGW method, consists of three-level schemes: the low-level, the middle-level and the high-level schemes. For the low-level schemes, the orthogonal matching pursuit combined with Gabor functions is proposed to provide an adaptive time-frequency decomposition of music signals. Compared with other algorithms for signal analysis, the proposed algorithm can achieve a higher spatial and temporal resolution and give a better interpret of the music signal structures. In the middle-level schemes, the Wigner distribution function is applied to obtain the time-frequency energy distribution of the results from the low-level schemes. High-level schemes are used to describe the modeling of audio features, the procedure of music mood classification. A classifier based on support vector machines is utilized to model the extracted features with the proposed technique regarding the emotion models. Several experiments are conducted with four datasets, and better results are achieved with the proposed method. In music mood classification experiments, music clips are classified into different kinds of mood clusters, and mean accuracy of 69.53 percent on our dataset can be achieved using the OMPGW method.

Social networks are critical in terms of information or malware propagation. However, how to contain the spreading of malware in social networks is still an open and challenging issue. In this paper, we propose a novel defending method through big data based influence modeling. We first establish a social interaction graph based on big data sets of the studied object. Based on the graph, we are able to measure direct influence of individuals by computing each node's strength, which includes the degree of the node and the total number of messages sent by each user to her friends. Then, we design an algorithm to construct influence spreading tree using the breadth first search strategy, and measure indirect influence of individuals by traversing the tree. We identify the top k influential nodes among all the nodes via the social influence strength, and propose an immunization algorithm to defend social networks against various attacks. The extensive experiments show that influence can spread easily in social networks, and the greater the influence of initial spread node is, the more impact it is on the malware propagation in social networks. The proposed method provides an effective solution to the prevention of malware or malicious messages propagation in social networks.

With the continued proliferation of smart mobile devices, the Quick Response (QR) code has become one of the most-used types of two-dimensional code in the world. Aiming at beautifying the visual-unpleasant appearance of QR codes, existing works have developed a series of techniques. However, these works still leave much to be desired, such as personalization, artistry, and robustness. To address these issues, in this paper, we propose a novel type of aesthetic QR codes, Stylized aEsthEtic (SEE) QR code , and a three-stage approach to automatically produce such robust style-oriented codes. Specifically, in the first stage, we propose a method to generate an optimized baseline aesthetic QR code, which reduces the visual contrast between the noise-like black/white modules and the blended image. In the second stage, to obtain an art style QR code, we tailor an appropriate neural style transformation network to endow the baseline aesthetic QR code with artistic elements. In the third stage, we design a module-based robustness-optimization mechanism to ensure the performance robust by balancing two competing terms: visual quality and readability. Extensive experiments demonstrate that the SEE QR code has high quality in terms of both visual appearance and robustness and also offers a greater variety of personalized choices to users.

Among various tools invented to help improve people's oral health, water flossers can achieve better performance than traditional and electronic toothbrushes, and are less harmful than dental floss, especially for those with orthodontic teeth or tooth implant surgeries. However, the water flossers available in the market serve no monitoring or recording functions that can help consumers clean their teeth in a more efficient way. To capture users' motions, this study develops a novel smart water flosser, installing an Inertial Measurement Unit (IMU) sensor on the handle of the flosser. We determine the motion cycle using signal processing techniques and extract a set of statistical characteristics from the data set. We then train and compare different machine learning models as classifiers to recognize the motions of the handle. We find that the Random Forest model achieves the best detection accuracy at 97% and 85% of the whole feature set and optimized set, respectively. Finally we implement an Android App that connects the smart water flosser with a Bluetooth module to show the washing area in real-time and record relevant information for further guidance.

Fast proliferation of mobile devices significantly promotes the development of mobile social networks. Users tend to interact with friends via multiple social networks. Multiple social networks identification is of great significance in terms of both attack and defense. Current methods either focus on the profile matching or network structure to re-identify a specific user. However, the accuracy are not satisfying with relative high error rate. In this paper, we propose a new Friendship learning-Based Identification (FBI) method to discriminate multiple pseudo identities of a real-world individual. We aim at providing potential attack mechanism to following privacy protection research. Firstly, we develop a new identification method based on friendship matching. Then, we implement a weighted mechanism which takes profile, network structure, and friendship into consideration. Furthermore, machine learning is leverage to further optimize the parameters and improve the accuracy. In addition, extensive experimental results show the superior of the FBI comparing to existing ones.

Public opinion events on the Internet are gaining more and more attention from the supervisory institutions for the possibility of malicious guide. Since the number of the events on the Internet is quite enormous, the process of supervision often costs a lot of manpower, which is contrary to the purposes and objectives of Sustainable Computing. However, most traditional methods for news recommendation are designed for netizens who do not have specific responsibilities like supervisory institutions. It is also difficult for supervisory institutions themselves to rate the public opinion articles, which is indispensable for recommendation. In this paper, a novel articles rating method based on tracking and matching (ARTM), is proposed for public opinion recommendation. The ARTM method can mine institution concerns from the browsing history and keep them updating automatically with the changing of institution attention. The processing flow of ARTM is as follows. Firstly, a set of institution concerns are established in terms of three aspects: fixed concerns, potential concerns and reading preferences. Then ratings of public opinion articles are computed by measuring the similarities between the vector of article keywords and the vector of institution concerns. Finally, articles are sorted by ratings and high- ranking articles are added to the recommendation list. In addition, the proposed rating algorithm and tracking algorithm can also be used as standalone modules for other services. In the end, comprehensive evaluation of the proposed method based on real data (78 supervisory institutions browsing history in one month) is made. Experimental results show that the proposed ARTM method can significantly improve recommendation efficiency.

With the rapid development of various online video sharing platforms, large numbers of videos are produced every day. Video affective content analysis has become an active research area in recent years, since emotion plays an important role in the classification and retrieval of videos. In this work, we explore to train very deep convolutional networks using ConvLSTM layers to add more expressive power for video affective content analysis models. Network-in-network principles, batch normalization, and convolution auto-encoder are applied to ensure the effectiveness of the model. Then an extended emotional representation model is used as an emotional annotation. In addition, we set up a database containing two thousand fragments to validate the effectiveness of the proposed model. Experimental results on the proposed data set show that deep learning approach based on ConvLSTM outperforms the traditional baseline and reaches the state-of-the-art system.

The Internet of Things (IoT) is an emerging technology and expected to provide solutions for various industrial fields. As a basic technology of the IoT, wireless sensor networks (WSNs) can be used to collect the required environment parameters for specific applications. Due to the resource limitation of sensor node and the open nature of wireless channel, security has become an enormous challenge in WSN. Authentication as a basic security service can be used to guarantee the legality of data access in WSN. Recently, Chang and Le proposed two authentication protocols for WSN for different security requirements. However, their protocol cannot provide proper mutual authentication and has other security and functionality defects. We present a three-factor user authentication protocol for WSN to remove the weaknesses of previous protocols. The security of the proposed protocol is analyzed, and the security, functionality and performance of our protocol are compared with other related protocols. The comparison results and simulation results by NS-3 show that the proposed protocol is robust and energy efficient for IoT applications.

Wireless sensor networks (WSNs) play an important role in the industrial Internet of Things (IIoT) and have been widely used in many industrial fields to gather data of monitoring area. However, due to the open nature of wireless channel and resource-constrained feature of sensor nodes, how to guarantee that the sensitive sensor data can only be accessed by a valid user becomes a key challenge in IIoT environment. Some user authentication protocols for WSNs have been proposed to address this issue. However, previous works more or less have their own weaknesses, such as not providing user anonymity and other ideal functions or being vulnerable to some attacks. To provide secure communication for IIoT, a user authentication protocol scheme with privacy protection for IIoT has been proposed. The security of the proposed scheme is proved under a random oracle model, and other security discussions show that the proposed protocol is robust to various attacks. Furthermore, the comparison results with other related protocols and the simulation by NS-3 show that the proposed protocol is secure and efficient for IIoT.

Smartphone built-in sensors are essential components of vehicle steering mode recognition. Related driver assistance systems (DAS) and abnormal driving behavior detection systems have been studied for many years. However, the existing solutions and systems simply collect sensor data with a fixed sliding window and fuse data from multiple sensors using simple thresholds to detect different driving behaviors. The weakness of these solutions can have an adverse impact on the energy consumption and computation complexity of power-limited devices such as smartphones, and may provide coarse-grained results. In this paper, we present a new method to reduce both the energy consumption and the computation complexity, and improve the recognition accuracy of vehicle steering patterns using the following three improvements: 1) a MultiWave filter is designed to replace the fixed sliding window, which is used to identify vehicle steering events; 2) a set of eight statistical sensor features reflecting the vehicle steering modes are identified by extracting statistical features from different sensors and different axes; 3) different machine learning methods are compared based on this feature set in order to improve classifier training (Decision Tree and Random Forest). Evaluation results based on real vehicle datasets show that our improved classifiers have high real-time recognition accuracy of the five most common steering modes: left/right turns, left/right lane changes and U-turns. We also took a simple on-road testing, in which both of the two models detected all the steering behaviors under low vehicle speed.

Running is one of the most important workouts to keep our body fit. This paper presents RunnerPal - a runner monitoring and advisory system by harmonizing the rhythms of breathing, heart beating and striding based on smart devices. RunnerPal is a convenient, biofeedback-based, automated music recommendation system, which utilizes Bluetooth headset, Apple Watch and smartphone to obtain body sensed data. To improve the accuracy of the detection, we propose a novel approach to calibrate the result by integrating ambient sensed data with a physiological model called Locomotor Respiratory Coupling (LRC), which indicates possible ratios between the striding and breathing frequencies. RunnerPal uses the sensed data and runner's contextual information to provide dynamic music suggestions to help the user achieve a target heart rate. We perform an empirical study to show the effect of music on heart rate and devise a Proportional Integral Differentiation Controller (PID - Controller) that recommends appropriate music to the user. RunnerPal has been validated by extensive experiments, and experimental results demonstrate that it can help runners achieve a target heart rate and maintain a stable running rhythm for indoor/outdoor running 91.6 percent of the time. In addition, RunnerPal can provide some advice to improve exercise effectiveness for runners.

A smartphone user's personal hotspot (pH) allows him to share a cellular connection to another (e.g., a traveler) in the vicinity, but such sharing consumes the limited data quota in his two-part tariff plan and may lead to an overage charge. This paper studies how to motivate such pH-enabled data-plan sharing between local users and travelers in the ever-growing roaming markets, and proposes pricing incentive for a data-plan buyer to reward surrounding pH sellers (if any). The pricing scheme practically takes into account the information uncertainty at the traveler side, including the random mobility and the sharing cost distribution of selfish local users who potentially share their pHs. Though the pricing optimization problem is non-convex, we show that there always exists a unique optimal price to a tradeoff between the successful sharing opportunity and the sharing price. We further generalize the optimal pricing to the case of heterogeneous selling pHs who have diverse data usage behaviors in the sharing cost distributions, and we show such diversity may or may not benefit the traveler. Lacking selfish pHs' information, the traveler's expected cost is higher than that under the complete information, but the gap diminishes as the pHs' spatial density increases. Finally, we analyze the challenging scenario that multiple travelers overlap for demanding data-plan sharing, by resorting to a near-optimal pricing scheme. We show that a traveler suffers as the travelers' spatial density increases.

In a rechargeable wireless sensor network, the data packets are generated by sensor nodes at a specific data rate, and transmitted to a base station. Moreover, the base station transfers power to the nodes by using Wireless Power Transfer (WPT) to extend their battery life. However, inadequately scheduling WPT and data collection causes some of the nodes to drain their battery and have their data buffer overflow, while the others waste their harvested energy, which is more than they need to transmit their packets. In this paper, we investigate a novel optimal scheduling strategy, called Scheduled WPT (SWPT), aiming to minimize data packet loss from a network of wireless powered sensor nodes by jointly considering the sensor nodes' energy consumption and data queue state information. The scheduling problem is formulated by a MDP model, assuming that the complete states of each sensor node are well known by the base station. This presents the best effort performance of the scheduling that can be collected in a wireless powered sensor network. The simulation results show that, in terms of network throughput and packet loss rate, the proposed scheduling model significantly improves the network performance.

With the pace of modern life quickening and increasing work stress, people don't have enough time to focus on their health and communicate with family members. The loneliness and chronic diseases (e.g., obesity, depression, diabetes, and dementia) have become more prevalent. In the paper, we propose SmartBuddy, a novel integrated mobile sensing and detecting system for monitoring people's family activities, which can motivate the user to do proper physical exercise and establish good relationships with family members for maintaining their physical and mental health. Specifically, SmartBuddy firstly uses smartphones and Apple Watches built-in sensors to obtain sensing data, such as the striding frequency and heart rate of the users, the sound of environment, etc. Secondly, SmartBuddy can accurately detect family activities including occurrence/duration of meal, cooking, TV viewing, conversations, in an unobtrusive manner based on sensed data. Thirdly, SmartBuddy will propose a personal plan to suggest the user doing some exercise and making continuous progress in the process of communicating with family members. We have fully implemented SmartBuddy on the Android platform and perform testbed experiments. The experimental results demonstrate that SmartBuddy is easy to use, accurate, and appropriate for family activities with the accuracy of 80% and the user satisfaction degree of 84.5%.

Nowadays, with increasing work stress and quick pace of modern life, people generally do not have enough time for exercising, however, they curiously pay much attention to the direct effect of exercising -calorie consumption. In this paper, we investigate several popular calorie consumption monitoring approaches and propose CCMS - a novel Calorie Consumption Monitoring System for exercising with least-squares calibration based on smartphones. Specifically, CCMS uses smartphone built-in sensors to collect the sensed data from accelerometer, barometer and GPS. With the sensed data, CCMS computes the calorie consumption based on the energy consumption formulas of American College of Sports Medicine. We apply an improved Naive Bayesian Classifier, which enables intelligent classification of several exercise types and achieves an average accuracy of 92.6% for determining the exercise types. We adopt the least-squares method to calibrate the result of calorie consumption and find that the method can increase the precision of CCMS up to 6%. We evaluate the performance of CCMS against other popular fitness applications, including Gudong and Jawbone UP3 which is a commercial device. The experimental results demonstrate that CCMS outperforms state-of-the-art calorie consumption monitoring systems in terms of measuring accuracy, with the average accuracy increase of about 5%.

The e-Health core concept includes Web usage in an integrated way with tools and services for healthcare. This definition improves access, efficiency, and clinical care quality process that are necessary for a service delivery improvement. Decision support systems (DSSs) belong to a plethora of e-Health concept dimensions. For these systems construction, it is important to find a reliable intelligent mechanism capable to identify diseases that can worsen the patient's clinical condition. Thus, this paper proposes the use of tree-based data mining (DM) techniques for the hypertensive disorders prediction in the risk gestation. It presents the modeling, performance evaluation, and comparison between the tree based classifiers ID3 and NBTree. The 5-fold cross-validation method realizes the performance comparison. Results show that the NBTree classifier obtained better performance, presenting F-measure 0.609, ROC area 0.753, and Kappa statistic 0.4658. This classifier can be a key to a smart system development capable to predict risk events in pregnancy. Therefore, DSSs are a leading solution for the reduction of both mother and fetal mortality.

In this paper, we propose a tooth brushing monitoring system based on acoustic inputs through an asymmetrical sound-field detector. This detector consists of a throat microphone and a Bluetooth earphone equipped on the user's neck and ear, respectively. This system can capture unique acoustic signals generated by the movement of the toothbrush on the surfaces of teeth via the detector. The throat microphone captures the brushing sound travelling through gums, bones, and muscles, which forms unique patterns with less attenuation than the sound travelling through the air. The Bluetooth earphone captures the brushing sound through the air. The tooth surface is divided into 16 parts for detection. By adopting machine learning models with the input of acoustic features from both time and frequency domains, we build a high accuracy detector to distinguish the brushing events happened at each of the 16 parts of the tooth surface. We employ Support Vector Machine (SVM), Hidden Markov Model (HMM), K-Means, C4.5 and Random Forest (RF) to evaluate the performance of our detection system. Experiments show that the RF model performs the best and achieves an average accuracy of 85.69\%. Based on the pre- trained model, we develop an Android-based APP to monitor the user's daily tooth brushing time and help the user form a good habit of tooth brushing.

A smartphone user's personal hotspot (pH) allows one to share cellular connection to another device nearby, but such sharing consumes the limited data quota in his or her two-part tariff plan and may lead to overage charge. This paper studies how to motivate such secondary data sharing via pHs for roaming markets, and proposes pricing incentive for a secondary data buyer (typically, a traveler) to opportunistically demand and reward pHs (if any) in the vicinity to reach a win-win situation. The pricing scheme practically takes into account the information uncertainty at the traveler side, including the random mobility and the sharing cost distribution of selfish local users who share pHs. Though the pricing optimization is non-convex problem, we show that there always exists a unique optimal price to tradeoff between the sharing opportunity and the sharing price, and can further extend the optimal pricing to the case of heterogeneous selling users/pHs who have diverse data usage behaviors. Lacking selfish pHs' information and cooperation, the traveler's expected cost is higher than that under the complete information, but the gap diminishes as the selfish pHs' spatial density increases. The traveler may or may not benefit from the diversity of pHs' data usage behaviors. Perhaps surprisingly, when the pHs' data usages are very diverse, the traveler's expected cost does not change with such diversity.

Automatic multi-document abstractive summarization system is used to summarize several documents into a short one with generated new sentences. Many of them are based on word-graph and ILP method, and lots of sentences are ignored because of the heavy computation load. To reduce computation and generate readable and informative summaries, we propose a novel abstractive multi-document summarization system based on chunk-graph (CG) and recurrent neural network language model (RNNLM). In our approach, A CG which is based on word-graph is constructed to organize all information in a sentence cluster, CG can reduce the size of graph and keep more semantic information than word-graph. We use beam search and character-level RNNLM to generate readable and informative summaries from the CG for each sentence cluster, RNNLM is a better model to evaluate sentence linguistic quality than n-gram language model. Experimental results show that our proposed system outperforms all baseline systems and reach the state-of-art systems, and the system with CG can generate better summaries than that with ordinary word-graph.

The task of identifying influential spreaders for various big data social network applications plays a crucial role in social networks, and lays the foundation for predictive or recommended applications. Though there are several kinds of methods for this task, most of these methods exploit global computing, and are time-consuming for large-scale social networks. In this paper, by combining the degree centrality with the law of universal gravitation in physics, we present a novel metric called Logarithm Gravity (LG) centrality to quantify the influence of nodes in large-scale social networks, which views the value of the degree centrality as mass for each node and regards the length of the shortest path between a pair of nodes as their distance. In our model, for each node, a local network is generated by obtaining all nodes, which are less than k-hop from it. Then the sum of mutual influence values between the node in question and all other nodes in each local network is figured out as its LG centrality index. Therefore, the complexity of our approach is scalable by adjusting the value of k with efficient local computation. We compare our LG centrality with k-shell, betweenness and degree centralities. Experimental evidence, which has been collected based on the SIR model with four real-world datasets, shows that our approach is more feasible and effective than other state-of-art methods in terms of infection ratios and computational complexity.

There is a growing trend towards attacks on database privacy due to great value of privacy information stored in big data set. Public's privacy are under threats as adversaries are continuously cracking their popular targets such as bank accounts. We find a fact that existing models such as K-anonymity, group records based on quasi-identifiers, which harms the data utility a lot. Motivated by this, we propose a sensitive attribute-based privacy model. Our model is the early work of grouping records based on sensitive attributes instead of quasi-identifiers which is popular in existing models. Random shuffle is used to maximize information entropy inside a group while the marginal distribution maintains the same before and after shuffling, therefore, our method maintains a better data utility than existing models. We have conducted extensive experiments which confirm that our model can achieve a satisfying privacy level without sacrificing data utility while guarantee a higher efficiency.

Taking part in family activities plays an important role in establishing good relationships with family members. It can solve the loneliness of elders, which related not only to their physical health, but also to the well-being of the whole family. In the paper, we propose FamilyPal, an effective system for detecting family activities, which can help users establish good relationship with family members. Specifically, FamilyPal firstly uses smartphones built-in sensors, such as GPS, accelerometer, microphone, gyroscope, and Wi-Fi to obtain the motion and location of users, the surrounding voice, etc. Secondly, with the sensed data, we propose an effective method based on Gaussian Mixtures Models (GMM) to detect family activities, including occurrence of meal, cooking, TV viewing, conversations, in an unobtrusive manner. Thirdly, we select appropriate sensors for classification to improve smartphones battery life. FamilyPal has been implemented on the Android platform and evaluation of the system with 10 subjects over one week shows that FamilyPal can accurately classify family activities with the average precision of 71%, the average recall of 73% and the F-measure of 71.99%.

An autopilot system includes several modules, and the software architecture has a variety of programs. As we all know, it is necessary that there exists one brand with a compatible sensor system till now, owing to complexity and variety of sensors before. In this paper, we apply (Robot Operating System) ROS-based distributed architecture. Deep learning methods also adopted by perception modules. Experimental results demonstrate that the system can reduce the dependence on the hardware effectively, and the sensor involved is convenient to achieve well the expected functionalities. The system adapts well to some specific driving scenes, relatively fixed and simple driving environment, such as the inner factories, bus lines, parks, highways, etc. This paper presents the case study of autopilot system based on ROS and deep learning, especially convolution neural network (CNN), from the perspective of system implementation. And we also introduce the algorithm and realization process including the core module of perception, decision, control and system management emphatically.

Underwater environments may vary gradually even when the occurrence of events is detected. Sensory data may follow a certain trend and are predictable during certain time durations. Taking these into consideration, a simple but practical data prediction mechanism is adopted for estimating sensory data and the geographical location of sensor nodes at sink nodes, and these data are synchronized with those sensed by underwater sensor nodes only when their variation is beyond a prespecified threshold. Leveraging these predicted data, the coverage and sources of potential events are identified by the sink node, and the evolution of these events is determined accordingly. Evaluation results show the applicability and energy-efficiency of this approach, especially when the variation of network environments follows certain and simple patterns.

Indoor localization based on smartphone can enhance user's experiences in indoor environments. Although some innovative solutions have been proposed in the past two decades, how to accurately and efficiently localize users in indoor environments is still a challenging problem. Traditional indoor positioning systems based on Wi-Fi fingerprints or dead reckoning suffer from the variation of Wi-Fi signals and the drift of dead reckoning problems, respectively. Crowdsourcing and ambient sensing stimulate new ways to improve existing localization systems' accuracy. Using human social factors to calibrate the accuracy of localization is practical and awarding. In this paper, we propose WAIPO, a collaborative indoor localization system with the fusion of Wi-Fi and magnetic fingerprints, image-matching, and people co-occurrence. Specifically, we could obtain the most likely top-n locations based on Wi-Fi fingerprints. We utilize the statistics of users' historical locations known by image-matching, for which we propose a photo-room matching algorithm, to reduce estimating areas. In order to further improve the accuracy of localization, we propose a co-occurrence and non-co-occurrence detection algorithm to detect users' spatial-temporal co-occurrence and determine users' locations with magnetic calibration. We have fully implemented WAIPO on the Android platform and perform testbed experiments. The experimental results demonstrate that WAIPO achieves an accuracy of 87.3% on average, which outperforms the state-of-the-art indoor localization systems.

User-generated reviews on the e-commerce site reflect consumers' sentiment about products, which can further direct consumers' purchasing behaviors and sellers' marketing strategies. In this paper, we propose a semi-supervised approach to mine the aspects of product discussed in Chinese online reviews and also the sentiments expressed in different aspects. We first apply the Latent Dirichlet Allocation model to discover multiaspect global topics of the product reviews, then extract the opinion short sentences based on sliding windows and pattern matching from context over the review text. The polarity of the associated sentiment is classified by the domain lexicon-based method. Finally the results are collected as features for the feedback for the machine learning method and applied in semantic orientation classification. The experiment results show that the novel method we proposed could help to discover multi-aspect fine-grained topics and associated sentiment, which helps to improve semantic orientation classification simultaneously.

This paper proposed a new kind of online education mode, MOOE (Massive Open Online Experiment). MOOE was produced based on the thoughts of the Internet thinking and the opening-and-sharing resources and was the organic combination of MOOC and experimental teaching. The main contributions were made: (1) Instructed MOOE's concept and characteristics. MOOE build a network laboratory with the advanced information technology to make experimenters complete all the experimental activities through the Internet without the limitation of time, space and resources. (2) Analyzed the similarities and differences between MOOE and MOOC. MOOE was the extension and expansion of MOOC, which realized the online learning of all the teaching activities including the theoretical courses and their related experiments. MOOE inherited the advantaged of MOOC, but paid more attention to the virtualization of experimental equipment, operations and contents. (3) Described the implement and application of MOOE by taking an experimental course titled “FPGA Based Multi-core Computing” as an example. The feasibility and advantage of MOOE were proved by the actual statistical data of the course.

The last decade has witnessed a dramatic growth of social networks, such as Twitter, Sina Microblog, etc. Messages/short texts on these platforms are generally of limited length, causing difficulties for machines to understand. Moreover, it is rarely possible for users to read and understand all the content due to the large quantity. So it is imperative to cluster and extract the viewpoints of these short texts. To solve this, the representation of a word is enriched with additional features from external, but it is demanding in terms of computational and time resources. In this paper, we proposed OnSeS, a novel short text summarization method which makes full use of word2vec to represent a word and utilizes neural network model to generate each word of the summary. OnSeS consists of three phrases: 1) clustering short texts using the K-means algorithm; 2) ranking content of each cluster by building a graph-based ranking model using BM25; 3) generating main point of each cluster with the help of neural machine translation model on the top ranked sentence. The experimental results reveal that our proposed fully data-driven approach outperforms state-of-the-art method.

Aggressive driving is the main cause of traffic accidents all over the world. Aggressive steering is the second leading cause of traffic accidents, just behind speeding. To detect aggressive steering and encourage drivers to cultivate better driving behaviors, automatic recognition of different vehicle steering modes is required so a decision can be taken on whether the behavior is aggressive or not. This paper investigates various patterns of turning, changing lanes and U-turns, and then design and implement a system termed MultiWave that utilizes the gyroscope of a smartphone to automatically detect vehicle steering patterns. MultiWave divides driving behaviors into different categories by analyzing the data collected by gyroscope sensors. Since it analyzes the gyroscopic sensor data of turning, changing lanes and U-turns, MultiWave is flexible and robust for most urban situations. The classification accuracy of MultiWave can reach averages of 92%, 76.5%, and 87% for vehicle turning, changing lanes and U-turn, respectively.

Scheduling of real-time tasks on a multi-core processor is challenging due to the execution time being a nondeterministic value. Through studying the relationship between task affinity and execution time, we propose an accelerated multi-core real-time scheduling algorithm (RM-λ) for periodic and dependent real-time tasks on a homogeneous multi-core processor based on acceleration between tasks to obtain a real-time scheduling scheme with less resource utilization. We adopt an acceleration factor matrix to represent the degree of affinity and develop a real-time scheduling model to find the best accelerated pair. The heterogeneous multi-core architectures can execute tasks by sharing their dependent data on L1 Cache. The results demonstrate our approach can loosens the schedulability constraints of RM (maximum improvement of 25%) so that an un-schedulable real-time tasks set on a single-core processor might be schedulable, and for those still hard to be scheduled, could be made schedulable on a multi-core processor.

In recent years, WiFi fingerprint-based localization has received much attention due to its deployment practicability. Although existing works show WiFi fingerprinting can achieve good localization accuracy, the experiments were conducted under their own testbeds within a small area and a short period. In this work, we investigate the impact of different indoor environmental factors, such as temporal and spatial similarity, on the performance of WiFi fingerprinting. We find that, WiFi fingerprinting is highly environment-sensitive. In an open space, it is quite challenging to find spatially varying but temporally stable signatures for adjacent reference locations. To address this issue, we propose a heatmap-based WiFi fingerprinting (called HMF) by utilizing layout construction as an additional input to improve WiFi fingerprint localization in open space environment. Our experimental results show, HMF can improve existing WiFi fingerprinting schemes like Radar and Horus by 28% and 80% in moderately open space, e.g., a wide corridor.

Online short texts of hot topics submitted to social media by users can provide valuable personal opinions, which are useful for service providers and individuals. However, it is difficult for readers to grasp the main opinions of massive short texts. In this paper, to cope with the summarization challenge of short texts, we proposed a novel approach, which makes full use of BM25 to weight each short text and syntactic parsing to generate important information of each opinion cluster. The approach also utilizes the feature pruning to reduce the dimensions of the vectors. We conduct our experiments on real datasets and evaluate the results by standard metrics and manual evaluation. The experimental results show that our proposed approach improves the accuracy when compared to the state-of-the-art method.

This paper proposes to detect event coverage and determines event sources. Generally, an appropriate sensor node is selected as the relay node for gathering and routing sensory data to sink node(s). When sensory data are collected at sink node(s), the event coverage is detected and represented as a weighted graph. Event sources are determined which correspond to the barycenters in this graph. Experiments show that this technique is more energy efficient, especially when the network topology is relatively steady.

Many-to-one data collection is a fundamental operation in wireless sensor networks (WSNs). To support long-term deployment of WSNs, sensor nodes normally operate at low-duty-cycles. However, the low-duty-cycle operation significantly reduces the communication chance between nodes. Consequently, the risk of data collisions significantly increases when multiple senders transmit packets to a receiver during its very short active period. Data collision not only results in wasted packet transmissions, but also incurs a large delivery latency. Under such conditions, collision-free medium access is more appealing than recovering after collision for low-duty-cycle WSNs. In this work, we propose an incast-collision-free data collection protocol, named iCore, to address the many-to-one collision problem in low-duty-cycle WSNs. iCore employs the dynamic forwarding technique and establishes a non-conflicting schedule for delay reduction. Specifically, we design efficient forwarder assignment and forwarding optimization algorithms that ensure low end-to-end latency under diverse data traffic types. Through comprehensive performance evaluations, we demonstrate that, compared with the state-of-the-art protocol, iCore effectively minimizes the end-to-end delay by 25% ∼ 57% and maintains high delivery ratio and energy efficiency for different many-to-one convergecast scenarios.

In this Demo, we present the design and implementation of a novel human localization and tracking system based on pyroelectric infrared(PIR) sensors. It is a new approach to setup several low-cost sensors to improve the simplicity of deployment. We first present a novel design in mechanical architecture, circuit board and sensors which enables our PIR sensor based system to detect the moving objects and track human beings in real time. Different from the previous work which only adopts binary output, we further explore the usage of analog signals from the sensors. This improvement is beneficial to accurate estimation of the distance between the device and human body. We have implemented several prototypes for evaluation and demonstration. According to the experimental results, our system can achieve an accuracy of 0.113m in an area of 12m × 6m.

Breathing plays an important role in the process of running. A stable and harmonic breathing rhythm can postpone runners' fatigue and help to improve their running performances. This paper presents a method that can detect runner's breathing frequency continuously. We utilize Bluetooth headset and smart phone to obtain sensed data, such as striding frequency and breathing frequency. Due to the interference of ambient noise, the detection will be inaccurate. In order to cope with this problem, we calibrate the detection result by leveraging a physiological model, called Locomotor Respiratory Coupling (LRC), which indicates possible ratios between the stride and breathing frequencies. Our method has been validated by extensive experiments and the experimental results indicate that it can accurately detect the breathing frequency for runners.

ABSTRACT With the development and maturation of the wireless communication technologies, the wireless sensor networks have been widely applied in different environments to acquire specific information. The wireless medical sensor networks (WMSNs), as a professional application of the wireless sensor networks in medicine, have attracted more and more attention because of its potential in improving the quality of healthcare services. Through the WMSNs, the parameters of patients' vital signs can be gathered from the sensor nodes equipped on the body of the patients and then can be accessed by the healthcare professionals by using a mobile device. By reason of the open feature of wireless communication, how to guarantee secure communication becomes an important issue. On the other hand, because the vital signs parameters are sensitive to the patients' health status and no one wants to reveal it to the others except the healthcare professionals, the protection of patients' privacy becomes another key issue for WMSNs applications. User authentication protocol with anonymity is the most basic and commonly used method to resolve the security and privacy issues of WMSNs. Recently, He et al. proposed an enhanced authentication protocol for healthcare applications using WMSNs to protect the security and privacy problems. However, we find that their scheme is incorrect in authentication and session key agreement phase. Besides, their scheme has no wrong password detection mechanism, which will not only waste the unnecessary computation and communication costs, but also may deduce the denial of service problem. In this paper, the biometric is introduced as the third authentication factor, and a new user anonymous authentication protocol based on WMSNs is designed so as to remove the drawbacks of the protocol of He et al. Compared with previous protocols, the new presented protocol enhances the security and also keeps the computation efficiency. Copyright © 2015 John Wiley & Sons, Ltd.

Abstract As the fast growth of multimedia information, the security of multimedia systems is becoming a rather important topic nowadays. Multimedia systems are often suffering attacks when users access the information and online services. Because of the excellent features of the biometric, many biometric‐based three‐factor remote user authentication schemes have been proposed to provide high level of security for different network‐based application systems. Recently, An pointed out the weaknesses of Das's three‐factor remote user authentication scheme and proposed an improved biometric‐based three‐factor remote user authentication scheme. An's scheme improves the security problems of previous schemes while keeping the efficiency. However, after detailed analysis, we find that An's scheme exists some weaknesses such as vulnerable to denial‐of‐service attack and forgery attack, cannot detect unauthorized login quickly, and does not provide session key agreement. In order to provide high level of security for multimedia systems, we design a robust three‐factor remote user authentication scheme with key agreement using elliptic curve cryptosystem. Copyright © 2014 John Wiley & Sons, Ltd.

The use of machine learning approaches for long-term hand tracking poses some major challenges such as attaining robustness to inconsistencies in lighting, scale and object appearances, background clutter, and total object occlusion/disappearance. To address these issues in this paper, we present a robust machine learning approach based on enhanced particle filter trackers. The inherent drawbacks associated with the particle filter approach, i.e., sample degeneration and sample impoverishment, are minimized by infusing the particle filter with the mean shift approach. Moreover, to instill our tracker with reacquisition ability, we propose a rotation invariant and efficient detection framework named beta histograms of oriented gradients. Our robust appearance model operates on the red, green, blue color histogram and our newly proposed rotation invariant noise compensated local binary patterns descriptor, which is a noise compensated, rotation invariant version of the local binary patterns descriptor. Through our experiments, we demonstrate that our proposed hand tracker performs favorably against state-of-the-art algorithms on numerous challenging video sequences of hand postures, and overcomes the largely unsolved problem of redetecting hands after they vanish and reappear into the frame.

Wireless sensor networks (WSNs) are integrated as a pillar of collaborative Internet of Things (IoT) technologies for the creation of pervasive smart environments. Generally, IoT end nodes (or WSN sensors) can be mobile or static. In this kind of hybrid WSNs, mobile sinks move to predetermined sink locations to gather data sensed by static sensors. Scheduling mobile sinks energy-efficiently while prolonging the network lifetime is a challenge. To remedy this issue, we propose a three-phase energy-balanced heuristic. Specifically, the network region is first divided into grid cells with the same geographical size. These grid cells are assigned to clusters through an algorithm inspired by the${{k}}$-dimensional tree algorithm, such that the energy consumption of each cluster is similar when gathering data. These clusters are adjusted by (de)allocating grid cells contained in these clusters, while considering the energy consumption of sink movement. Consequently, the energy to be consumed in each cluster is approximately balanced considering the energy consumption of both data gathering and sink movement. Experimental evaluation shows that this technique can generate an optimal grid cell division within a limited time of iterations and prolong the network lifetime.

Reliable flooding in wireless sensor networks (WSNs) is desirable for a broad range of applications and network operations. However, relatively little work has been done for reliable flooding in low-duty-cycle WSNs with unreliable wireless links. It is a challenging problem to efficiently ensure 100% flooding coverage considering the combined effects of low-duty-cycle operation and unreliable wireless transmission. In this paper, we propose a novel dynamic switching-based reliable flooding (DSRF) framework, which is designed as an enhancement layer to provide efficient and reliable delivery for a variety of existing flooding tree structures in low-duty-cycle WSNs. The key novelty of DSRF lies in the dynamic switching decision making when encountering a transmission failure, where a flooding tree structure is dynamically adjusted based on the packet reception results for energy saving and delay reduction. DSRF distinguishes itself from the existing works in that it explores both poor links and good links on demand. In addition, we define the optimal wakeup schedule-ranking problem in order to maximize the switching gain in DSRF. We prove the NP-completeness of this problem and present a heuristic algorithm with a low computational complexity. Through comprehensive performance comparisons, including the simulation of large-scale scenarios and small-scale experiments on a WSN testbed, we demonstrate that compared with the flooding protocol without DSRF enhancement, the DSRF effectively reduces the flooding delay and the total number of packet transmission by 12%' 25% and 10%' 15%, respectively. Remarkably, the achieved performance is close to the theoretical lower bound.

Flooding, by which the sink node broadcasts messages to the entire network, is an important and common operation in wireless sensor networks (WSNs). The emerging synchronous duty-cycled wakeup schedules together with the unreliable communication of WSNs post new challenges for efficient broadcast protocol design, and the existing methods are not appropriate to address this problem. This paper proposes a delay-aware energy-optimized flooding algorithm (DEF) tailored for synchronous duty-cycled WSNs, which can act as an enhanced scheme for most flooding trees. DEF globally adjusts a constructed flooding tree, to maximize the energy efficiency improvement while following the delay constraint. To this end, we first remodel the flooding problem in the new context mathematically. Then, we introduce a routing metric, which fully utilizes the features of synchronous duty-cycled WSNs for energy optimization, and design an effective delay-aware tree adjusting approach. Extensive evaluation results demonstrate that DEF could save considerable energy, while the flooding delay keeps unchanged or even decreases slightly. In addition, a modified minimum spanning tree (MMST) is proposed to indicate the approximate energy lower bound. Compared with MMST, DEF achieves comparable energy efficiency and better latency performance.

Digital Living Network Alliance (DLNA) puts forward an interoperable architecture of home network equipment to implement the simple and seamless interoperability between household appliances, mobile devices and computers, so as to enhance and enrich users' experience. Because DLNA is designed to implement the sharing of multimedia resources between devices in a family environment, it does not support cross-LAN accessing. Currently, the cross-network cooperative work between smart devices has become a hot issue. To overcome this problem, this paper presents CLMRS - a cross-LAN accessing solution of C/S architecture based on DLNA technology. CLMRS can be used to implement cross-network media resources sharing by designing DLNA gateway/router of applicaion-level. We use DLNA gateway as a agent of LAN which is designed to transpond communication messages and to redirect address. To achieve cross-LAN media resources sharing, CLMRS adapts DLNA router as a transfer server to transpond communication data and to manage users access right. We also optimize the transmission path of media Steam to reduce the pressure of DLNA router. In order to protect user privacy, CLMRS sets a family account to limit the access right and proposes an efficient and secure authentication mechanism with anonymity. Our design and theoretical model are validated via implementing an instance of DLNA cross-network communication. Experimental results show that the approach is practical and can protect user privacy effectively.

There has been a significant body of work on evaluating node criticality in information networks. However, most of the existing works are developed for static networks and are not applicable to dynamic settings where connectivities among nodes change frequently over time. In this paper, we treat an opportunistic mobile social network as a time-evolving, dynamic graph, and propose a scheme to ascertain the information dissemination capability for each node based on its contact history. In particular, we analyze the node importance in spreading or forwarding real-time messages which are assumed to become less important or even stale over time. To this end, we take a dynamic walk counting approach to calculate all possible temporal-spatial routes associated with each node, by using the down-weighting method. Since the age of a message increases with time, the old walks are discounted to represent the fading influence on the target node. Extensive experiments are conducted based on 4 real-world trace datasets, and the results show that, our analytical result is effective at ranking the node criticality in disseminating or acquiring real-time messages in opportunistic mobile social networks.

WiFi fingerprint-based indoor localization techniques have been proposed and widely used in recent years. Most solutions need a site survey to collect fingerprints from interested locations to construct the fingerprint database. However, the site survey is labor-intensive and time-consuming. To overcome this shortcoming, we record user motions as well as WiFi signals without the active participation of the users to construct the fingerprint database, in place of the previous site survey. In this paper, we develop an indoor localization system called WicLoc, which is based on WiFi fingerprinting and crowdsourcing. We design a fingerprint model to form fingerprints of each location of interest after fingerprint collection. We propose a weighted KNN (K-Nearest Neighbor) algorithm to assign different weights to APs and achieve room-level localization. To obtain the absolute coordinate of users, we design a novel MDS (Multi-Dimensional Scaling) algorithm called MDS-C (Multi-Dimensional Scaling with Calibrations) to calculate coordinates of interested locations in the corridor and rooms, where anchor points are used to calibrate absolute coordinates of users. Experimental results show that our system can achieve a competitive localization accuracy compared with state-of-the-art WiFi fingerprint-based methods while avoiding the labor-intensive site survey.

Overlapping community detection algorithm research is one of hot topics in current social network analysis. In this paper, we applied the idea of weighted label propagation to overlapping community detection algorithm design, and propose a weighted label propagation algorithm (WLPA). Moreover, in order to evaluate the performance results of various overlapping community detection algorithms, we put forward a series of evaluation criteria based on error distribution curve of overlapping vertices. The experiment results show that the algorithm has a faster speed and better community detection results, and the evaluation criteria is in line with the inherent characteristics of the social network overlapping community structure.

Prior studies show that repairing partially corrupted packets, instead of retransmitting them in their entirety, holds potential in improving the performance of 802.11 networks. However, the efficiency of existing packet recovery approaches is severely limited by various overhead associated to redundant transmission and repeated channel contention. In this paper, we propose CodeRepair, a practical coding-based protocol that recovers partially corrupted 802.11 packets without these pains. The design of CodeRepair is based on two novel ideas. First, CodeRepair pushes the limit of 802.11 PHY to piggyback parities in the padded bits of OFDM, obviating the need of transmitting extra information for error correction. Second, CodeRepair corrects errors at the PHY layer, which is significantly more efficient than traditional link-layer approaches. This is due to the fact that a single coded bit usually affects the decoding of a group of data bits in 802.11 convolutional code. As a result, CodeRepair can salvage a partially corrupted packet by correcting a small number of erroneous coded bits using the padded parities. To reduce computational cost of error recovery, CodeRepair employs single parity code for correcting coded bit errors. We propose several techniques to augment the error correcting capability of single parity code without compromising its computation efficiency. Our evaluation shows that CodeRepair recovers an average of 34% partially corrupted packets, and improves the end-to-end link goodput by 59% on lossy 802.11 links.

This paper presents VINCE - a novel visible light sensing design for smartphone-based Near Field Communication (NFC) systems. VINCE encodes information as different brightness levels of smartphone screens, while receivers capture the light signal via light sensors. In contrast to RF technologies, the direction and distance of such a Visible Light Communication (VLC) link can be easily controlled, preserving communication privacy and security. As a result, VINCE can be used in a wide range of NFC applications such as contactless payments and device pairing. We experimentally profile the impact of screen brightness levels and refresh rates of smartphones, and then use the results to guide the design of light intensity encoding scheme of VINCE. We adopt several signal processing techniques and empirically derive a model to deal with the significant variation of received light intensity caused by noises and low screen refresh rates. To improve the communication reliability, VINCE adopts a feedback-based retransmission scheme, and dynamically adjusts the number of encoding brightness levels based on the current light channel condition. We also derive an analytical model that characterizes the relation among the distance, SNR (Signal to Noise Ratio), and BER (Bit Error Rate) of VINCE. Our design and theoretical model are validated via extensive evaluations using a hardware implementation of VINCE on Android smartphones and the Arduino platform.

With the rapid popularization of smartphones and tablets, there are thousands of applications based on mobile social networks. The big data from these networks provide a huge potential to shed light on the mobility patterns of users. These big data enable a deeper understanding of users' preferences and behaviors and will help us mine users' friendship in both physical and digital worlds. In this paper, we firstly divide user mobility patterns into different categories to portray the characteristics of user encounter more precisely. Then, with combining proximity data from bluetooth devices and location data from cellular towers, we introduce a set of spatial-temporal features, including the encounter entropy, which measures the probability of encounters between different mobile users. Using these spatial-temporal features, we provide a novel model to infer user friendship by analyzing the social context of users and their encounters. To address the class imbalance problem in the dataset and improve the prediction accuracy of friendship, we employ the sampling method and evaluate our model with three different classifiers. The experimental results show that our encounter entropy feature has a striking effect to infer user friendship, and our model based on these spatial-temporal features can achieve pretty good accuracy in predicting friendship over real human mobility traces without privacy-sensitive information disclosure.

Participatory sensing enables to collect a vast amount of data from the crowd by allowing a wide variety of sources to contribute data. However, the openness of participatory sensing exposes the system to malicious and erroneous participations, inevitably resulting in poor data quality. This brings forth the important issues of false data detection and correction in participatory sensing. Furthermore, data collected by participants normally include considerable missing values, which poses challenges for accurate false data detection. In this work, we propose DECO, a general framework to detect false values for participatory sensing in the presence of missing data. By applying a tailored spatio-temporal compressive sensing technique, DECO is able to accurately detect the false data and estimate both false and missing values for data correction. We validate our design through an experimental case study.

Radio duty cycling is a commonly employed mechanism to support long-term sustainable operations of WSNs. Combined with the effect of unreliable wireless links, many challenges arise for ensuring delay bounded data delivery with reliability constraint. However, research on energy-efficient data forwarding with statistical delay bound in duty-cycled WSNs still remains unaddressed. This paper proposes EDGE, a novel opportunistic forwarding technique tailored for duty-cycled WSNs with unreliable wireless links. The key idea is to exploit the available path diversity to minimize the transmission cost while providing statistical delay guarantees. Delay quantiles are derived at each node in a distributed manner and are used as the guidelines in forwarding decision making, so that an early arriving packet will be opportunistically switched to the energy-optimal path for communication cost minimization. Comprehensive evaluation results show that EDGE effectively reduces the transmission cost with statistical delay guarantees under various network settings.

Environment reconstruction is to rebuild the physical environment in the cyberspace using the sensory data collected by sensor networks, which is a fundamental method for human to understand the physical world in depth. A lot of basic scientific work such as nature discovery and organic evolution heavily relies on the environment reconstruction. However, gathering large amount of environmental data costs huge energy and storage space. The shortage of energy and storage resources has become a major problem in sensor networks for environment reconstruction applications. Motivated by exploiting the inherent feature of environmental data, in this paper, we design a novel data gathering protocol based on compressive sensing theory and time series analysis to further improve the resource efficiency. This protocol adapts the duty cycle and sensing probability of every sensor node according to the dynamic environment, which cannot only guarantee the reconstruction accuracy, but also save energy and storage resources. We implement the proposed protocol on a 51-node testbed and conduct the simulations based on three real datasets from Intel Indoor, GreenOrbs and Ocean Sense projects. Both the experiment and simulation performances demonstrate that our method significantly outperforms the conventional methods in terms of resource efficiency and reconstruction accuracy.

In existing WiFi-based localization methods, smart mobile devices consume quite a lot of power as WiFi interfaces need to be used for frequent AP scanning during the localization process. In this work, we design an energy-efficient indoor localization system called ZigBee assisted indoor localization (ZIL) based on WiFi fingerprints via ZigBee interference signatures. ZIL uses ZigBee interfaces to collect mixed WiFi signals, which include non-periodic WiFi data and periodic beacon signals. However, WiFi APs cannot be identified from these WiFi signals by ZigBee interfaces directly. To address this issue, we propose a method for detecting WiFi APs to form WiFi fingerprints from the signals collected by ZigBee interfaces. We propose a novel fingerprint matching algorithm to align a pair of fingerprints effectively. To improve the localization accuracy, we design the K-nearest neighbor (KNN) method with three different weighted distances and find that the KNN algorithm with the Manhattan distance performs best. Experiments show that ZIL can achieve the localization accuracy of 87%, which is competitive compared to state-of-the-art WiFi fingerprint-based approaches, and save energy by 68% on average compared to the approach based on WiFi interface.

Abstract With their proliferation and increasing capabilities, mobile devices with local wireless interfaces can be organized into delay‐tolerant networks (DTNs) that exploit communication opportunities arising out of the movement of their users. As the mobile devices are usually carried by people, these DTNs can also be viewed as social networks. Unfortunately, most existing routing algorithms for DTNs rely on relatively simple mobility models that rarely consider these social network characteristics, and therefore, the mobility models in these algorithms cannot accurately describe users’ real mobility traces. In this paper, we propose two predict and spread (PreS) message routing algorithms for DTNs. We employ an adapted Markov chain to model a node's mobility pattern and capture its social characteristics. A comparison with state‐of‐the‐art algorithms demonstrates that PreS can yield better performance in terms of delivery ratio and delivery latency, and it can provide a comparable performance with the epidemic routing algorithm with lower resource consumption. Copyright © 2013 John Wiley & Sons, Ltd.

The number of Non-communicable disease (NCD) patients in China is growing rapidly, which is far beyond the capacity of the national health and social security system. Community health stations do not have enough doctors to take care of their patients in traditional ways. In order to establish a bridge between doctors and patients, we propose eBPlatform, which is an information system based on the Internet of Things (IoT) technology for homecare of the NCD patients. The eBox is a sensor which can be deployed in the patient's home for blood pressure measurement, blood sugar measurement and ECG signals collection. Some services are running on the remote server, which can receive the samples, filter and analyze the ECG signals. The uploaded data will be pushed to a web portal, with which doctors provide treatments online. The system requirements, design and implementation of hardware and software are discussed respectively. Finally, we investigate a case study with 50 NCD patients for half a year in Beijing. The results show that eBPlatform can increase the efficiency of the doctor and make a big progress to eliminate the numerical imbalance between community medical practitioners and NCD patients.

Identifying the most influential spreaders in social networks has many practical applications. The existing methods for the purpose are either too time-consuming for dynamic large-scale networks, such as betweenness centrality, closeness centrality, eigenvector centrality and Katz centrality, or do not consider the network topology, such as degree centrality. To design an effective method to identify the most influential nodes in a network, we propose a novel metric, k-hop centrality which is a generalization of degree centrality. The k-hop index is the summation of the number n(i) of nodes within k-hop distance from the node in question, attenuated by 1/αi, for 1 ≤ i ≤ k (α is the average degree of nodes in the network). It is calculated in a localized manner and is complexity-scalable by adjusting the value of k, thus suitable for dynamically changing, large social networks. We adopt the Susceptible Infected Recovered (SIR) model to evaluate the performance of k-hop centrality over four real datasets of complex networks, and experimental results show that our method outperforms state-of-the-art methods in this field in terms of both infection ratios (spreading influence) and computational complexity. Our work sheds some light on designing efficient spreading strategies for complex networks.

Among the current clustering algorithms of complex networks, Laplacian-based spectral clustering algorithms have the advantage of rigorous mathematical basis and high accuracy. However, their applications are limited due to their dependence on prior knowledge, such as the number of clusters. For most of application scenarios, it is hard to obtain the number of clusters beforehand. To address this problem, we propose a novel clustering algorithm - Jordan-Form of Laplacian-Matrix based Clustering algorithm (JLMC). In JLMC, we propose a model to calculate the number (n) of clusters in a complex network based on the Jordan-Form of its corresponding Laplacian matrix. JLMC clusters the network into n clusters by using our proposed modularity density function (P function). We conduct extensive experiments over real and synthetic data, and the experimental results reveal that JLMC can accurately obtain the number of clusters in a complex network, and outperforms Fast-Newman algorithm and Girvan-Newman algorithm in terms of clustering accuracy and time complexity.

We investigate the group growth in online social networks, by analyzing six different user groups (two million users in total) in Douban Network. The size and longevity of posts in the Douban dataset demonstrate a power-law distribution with exponential cutoff and heavy tail, respectively. The frequency of user interactions follows a two-stage power-law distribution, which can distinguish different types of users. The growth of the number of users and the number of posts/replies generated by the users in a given and same time period, in each group, follow an exponential pattern at the initial stage and oscillate dramatically during the rest of the processes. The number of posts/replies has a power-law relation with the number of active users within a period of time. We propose an empirical growth model, Twisted Growth (TG), to portray the relation between the number of users and the amount of the contents they generated. The model derives equations based on the historical data for deciding coefficients, and the assumtion that the contents in one group will attract new users to join, which will lead to growth of users. Further, the newcomers together with original users will create new contents. We validate our TG model through theoretical analysis and simulations over real datasets.

A novel near field communication system for s-martphones based on visible light (NECAS) is presented in this paper. NECAS encodes information using combinations of different colors (red, green and blue) and their intensity levels of a smartphone screen, while the receiver captures the light signal via color sensors. In our scheme, a smartphone screen is segmented into four sub-blocks, and each of them can be used to encode information independently. NECAS can be used in a range of Near Field Communication applications such as contactless payments, electronic ticket checking and device paring. We design the color encoding scheme by experimentally profiling the interference of three color channels and the refresh rate of smartphones. We implement a joint decoding algorithm to deal with the significant interference among color channels. With a feedback channel, the retransmission scheme is adopted to improve communication reliability, and the system can dynamically adjust the number of combinations of different colors and intensity levels to fit the communication environment. Our design is validated via extensive experiments using a hardware implementation on Android smartphones and an Arduino platform.

ABSTRACT There are some biometrics‐based three‐factor remote user authentication schemes proposed by researchers for ensure high security features for network‐based application systems. Recently, Das pointed out the security flaws of Li and Hwang's three‐factor remote user authentication scheme, and proposed an enhanced biometrics‐based three‐factor remote user authentication scheme. Das's scheme overcomes the defects of Li and Hwang's scheme, and maintains the advantages of Li and Hwang's scheme at the same time. However, after detailed analysis, we find that Das's scheme remains vulnerable to forgery attack and stolen smart card attack; at the same time, Das's scheme cannot provide the session key agreement after the mutual authentication. To provide more security features, we design a three‐factor remote user authentication scheme with key agreement using biometrics. Copyright © 2013 John Wiley & Sons, Ltd.

ABSTRACT User authentication and privacy protection are important issues for wireless and mobile communication systems such as GSM, 3G, and 4G wireless networks. Recently, Yoon et al. proposed a user‐friendly authentication scheme with anonymity for wireless communications. However, in this paper, we show that user anonymity of their scheme is not achieved under the eavesdropping attack and their scheme is not fair in the key agreement. In order to ensure security authentication and protect user anonymity for wireless communications, we propose a novel user authentication scheme with anonymity based on elliptic curve cryptosystem, which can resist various known types of attacks and is more practical for wireless and mobile communications. Copyright © 2012 John Wiley & Sons, Ltd.

Providing reliable and efficient communication under fading channels is one of the major technical challenges in wireless sensor networks (WSNs), especially in industrial WSNs (IWSNs) with dynamic and harsh environments. In this work, we present the Reliable Reactive Routing Enhancement (R3E) to increase the resilience to link dynamics for WSNs/IWSNs. R3E is designed to enhance existing reactive routing protocols to provide reliable and energy-efficient packet delivery against the unreliable wireless links by utilizing the local path diversity. Specifically, we introduce a biased backoff scheme during the route-discovery phase to find a robust guide path, which can provide more cooperative forwarding opportunities. Along this guide path, data packets are greedily progressed toward the destination through nodes' cooperation without utilizing the location information. Through extensive simulations, we demonstrate that compared to other protocols, R3E remarkably improves the packet delivery ratio, while maintaining high energy efficiency and low delivery latency.

As a key approach to achieve energy efficiency in sensor networks, sensing coverage has been studied extensively in the literature. Researchers have designed many coverage protocols to provide various kinds of service guarantees on the network lifetime, coverage ratio and detection delay. While these protocols are effective, they are not flexible enough to meet multiple design goals simultaneously. In this paper, we propose a unified sensing coverage architecture for duty cycled wireless sensor networks, called uSense, which features three novel ideas: Asymmetric Architecture, Generic Switching and Global Scheduling. We propose asymmetric architecture based on the conceptual separation of switching from scheduling. Switching is efficiently supported in sensor nodes, while scheduling is done in a separated computational entity, where multiple scheduling algorithms are supported. As an instance, we propose a two-level global coverage algorithm, called uScan. At the first level, coverage is scheduled to activate different portions of an area. We propose an optimal scheduling algorithm to minimize area breach. At the second level, sets of nodes are selected to cover active portions. Importantly, we show the feasibility to obtain optimal set-cover results in linear time if the layout of areas satisfies certain conditions. Through extensive testbed and simulation evaluations, we demonstrate that uSense is a promising architecture to support flexible and efficient coverage in sensor networks.

QoS routing is an important research issue in wireless sensor networks (WSNs), especially for mission-critical monitoring and surveillance systems which requires timely and reliable data delivery. Existing work exploits multipath routing to guarantee both reliability and delay QoS constraints in WSNs. However, the multipath routing approach suffers from a significant energy cost. In this work, we exploit the geographic opportunistic routing (GOR) for QoS provisioning with both end-to-end reliability and delay constraints in WSNs. Existing GOR protocols are not efficient for QoS provisioning in WSNs, in terms of the energy efficiency and computation delay at each hop. To improve the efficiency of QoS routing in WSNs, we define the problem of efficient GOR for multiconstrained QoS provisioning in WSNs, which can be formulated as a multiobjective multiconstraint optimization problem. Based on the analysis and observations of different routing metrics in GOR, we then propose an Efficient QoS-aware GOR (EQGOR) protocol for QoS provisioning in WSNs. EQGOR selects and prioritizes the forwarding candidate set in an efficient manner, which is suitable for WSNs in respect of energy efficiency, latency, and time complexity. We comprehensively evaluate EQGOR by comparing it with the multipath routing approach and other baseline protocols through ns-2 simulation and evaluate its time complexity through measurement on the MicaZ node. Evaluation results demonstrate the effectiveness of the GOR approach for QoS provisioning in WSNs. EQGOR significantly improves both the end-to-end energy efficiency and latency, and it is characterized by the low time complexity.

The trade-off between system performance and energy efficiency (service time) is critical for battery-based embedded systems. Most of the previous work focuses on saving energy in a deterministic way by taking the average or worst scenario into account. However, such deterministic approaches usually are inappropriate in modeling energy consumption because of uncertainties in conditional instructions on processors and time-varying external environments (e.g., fluctuant network bandwidth and different user inputs). By adopting a probabilistic approach, this paper proposes a model and a set of algorithms to address the Processor and Voltage Assignment with Probability (PVAP) problem of data-dependent aperiodic tasks in real-time embedded systems, ensuring that all the tasks can be done under the time constraint with a guaranteed probability. We adopt a task DAG (Directed Acyclic Graph) to model the PVAP problem. We first use a processor scheduling algorithm to map the task DAG onto a set of voltage-variable processors, and then use our dynamic programming algorithm to assign a proper voltage to each task. Finally, to escape from local optima, a local search with restarts searches the optimal solution from candidate solutions by updating the objective function, until the stop criteria are reached or a time bound is elapsed. The experimental results demonstrate that for probability 1.0, our approach yields slightly better results than the well-known algorithms like ASAP/ALAP (As Soon As Possible/As Late As Possible) and ILP (Integer Linear Programming) with/without DVS (Dynamic Voltage Scaling). However, for probabilities 0.8 and 0.9, our approach significantly outperforms those algorithms (maximum improvement of 50.3 percent).

Nowadays, networks play an indispensable role in social life, and social networks have become a new advertising medium for offline activities. Previous studies of information diffusion or behavior spread over social networks have mostly focused on diffusion models and analysis of virtual interaction between online users, and very few of them focus on the propagation of real world activities in these social networks. To address this problem, we use data obtained from the Chinese largest recommendation social network - Douban, and study how the offline activities spread from one user to another through Douban. By using cascading subgraphs and diffusion trees, we break a whole cascade into local subgraphs. After analyzing the activities of about 1.47 million users, we observe the statistical and topological characteristics of these local cascading subgraphs. Next, we find the size and degree distributions of these cascading subgraphs and several common patterns of topology of local cascades. Moreover, we also have some other interesting discoveries, like the relation between the number of initial adopters and the final cascade size, and the underlying influences driving user behaviors. Finally, we propose a diffusion model that can generate information cascades that follow the patterns we have observed, and validate it by empirical analysis.

The Vehicular Ad hoc Networks (VANETs) have been recently introduced to provide high-speed Internet access to vehicles by deploying 802.11 enhanced Roadside Units (RSUs) along roads. However, few content downloading oriented RSU deployment strategies have been proposed. In this paper, we propose a new RSU deployment strategy for content downloading in VANETs. The encounters between vehicles and RSUs are modeled as a time continuous homogeneous Markov chain. The optimal inter-meeting time between vehicles and RSUs is analyzed based on the encounter model. Then, the road network is modeled as a weighted undirected graph, and a RSU deployment algorithm is designed based on the depth-first traversal algorithm for edges of a graph. Simulation results show that the proposed RSU deployment algorithm can satisfy the file downloading service requirements with the lowest RSU deployment cost.

With the proliferation of high-end mobile devices that feature wireless interfaces, many promising applications are enabled in opportunistic networks. In contrary to traditional networks, opportunistic networks utilize the mobility of nodes to relay messages in a store-carry-forward paradigm. Thus, the relay process in opportunistic networks faces several practical challenges in terms of delay and delivery ratio. In this paper, we propose a novel P2P Query algorithm based on Betweenness Centrality Forwarding (PQBCF), for opportunistic networks. PQBCF adopts a forwarding metric called Betweenness Centrality (BC), which is borrowed from social networks, to quantify the active degree of nodes in the networks. In PQBCF, nodes with higher BC are preferable to serve as relays, leading to higher inquiry success ratio and lower inquiry delay. A comparison with the state-of-the-art algorithms reveals that PQBCF can provide better performance on both the query success ratio and query delay, and approaches the performance of Flooding with much less resource consumption.

Partitioning and offloading some parts of mobile applications onto remote servers is a promising approach to extend the battery life of mobile devices. However, since available network bandwidths vary in a wireless environment, static partitionings proposed by previous works with a fixed bandwidth assumption are unsuitable for mobile platforms, while dynamic partitionings result in high overhead due to continuously partitioning. Targeting this problem, we propose a novel partitioning scheme taking the bandwidth as a variable to improve static partitioning and avoid high costs of dynamical partitioning. Based on the application Object Relation Graph, we propose a partitioning optimization model and two bandwidth-adaptive partitioning algorithms: Branch-and-Bound based Application Partitioning (BBAP) and Min-Cut based Greedy Application Partitioning (MCGAP). BBAP is suitable for obtaining the optimal partitionings for small applications, while MCGAP is applicable to large-scale applications by quickly obtaining suboptimal solutions. Experimental results demonstrate that both algorithms can adapt to bandwidth fluctuations well, and significantly reduce the execution time and energy consumption by optimally distributing components between mobile devices and servers.

Reliable flooding in wireless sensor networks (WSNs) is desirable for a broad range of applications and network operations, and has been extensively investigated. However, relatively little work has been done for reliable flooding in lowduty-cycle WSNs with unreliable wireless links. It is a challenging problem to efficiently ensure 100% flooding coverage considering the combined effects of low-duty-cycle operation and unreliable wireless transmission. In this work, we propose a novel dynamic switching-based reliable flooding (DSRF) framework, which is designed as an enhancement layer to provide efficient and reliable delivery for a variety of existing flooding tree structures in lowduty-cycle WSNs. The key novelty of DSRF lies in the dynamic switching decision making when encountering a transmission failure, where a flooding tree structure is dynamically adjusted based on the packet reception results for energy saving and delay reduction. DSRF is distinctive from existing works in that it explores both poor links and good links on demand. Through comprehensive performance comparisons, we demonstrate that, compared with the flooding protocol without DSRF enhancement, DSRF effectively reduces the flooding delay and the total number of packet transmission by 12% 25% and 10% 15%, respectively. Remarkably, the achieved performance is close to the theoretical lower bound.

Indoor localization becomes increasingly important as context-aware applications gain popularity in mobile users. A promising approach for indoor localization is to leverage the pervasive WiFi infrastructure via fingerprinting-based inference. However, a WiFi device must frequently scan for WiFi signals during localization, leading to high power consumption. Moreover, switching to the scanning mode introduces inevitable disruptions to data communication of WiFi interface. This paper presents a new indoor localization system called ZiFind that exploits the cross-technology interference in the unlicensed 2.4 GHz frequency spectrum. ZiFind utilizes low-power ZigBee interface to collect WiFi interference signals and adopts digital signal processing techniques to extract unique signatures as fingerprints for localization. To deal with the noise in the fingerprints, we design a new learning algorithm called R-KNN that can improve the accuracy of localization by assigning different weights to fingerprint features according to their importance. We implement ZiFind on TelosB motes and evaluate its performance through extensive experiments in a 16,000 ft2office building floor consisting of 28 rooms. Our results show that ZiFind leads to significant power saving compared with existing approaches based on WiFi interface, and yields satisfactory localization accuracy in a range of realistic settings.

This paper focuses on studying the sensor nodes deployment towards accurate data fusion for ship movement controlling. Furthermore, this study provides a node deployment layout with better measurement accuracy, which is surprisedly different from the layout that we originally predicted.

Location-based services, such as targeted advertisement, geo-social networking and emergency services, are becoming increasingly popular for mobile applications. While GPS provides accurate outdoor locations, accurate indoor localization schemes still require either additional infrastructure support (e.g., ranging devices) or extensive training before system deployment (e.g., WiFi signal fingerprinting). In order to help existing localization systems to overcome their limitations or to further improve their accuracy, we propose Social-Loc, a middleware that takes the potential locations for individual users, which is estimated by any underlying indoor localization system as input and exploits both social encounter and non-encounter events to cooperatively calibrate the estimation errors. We have fully implemented Social-Loc on the Android platform and demonstrated its performance on two underlying indoor localization systems: Dead-reckoning and WiFi fingerprint. Experiment results show that Social-Loc improves user's localization accuracy of WiFi fingerprint and dead-reckoning by at least 22% and 37%, respectively. Large-scale simulation results indicate Social-Loc is scalable, provides good accuracy for a long duration of time, and is robust against measurement errors.

Although social-aware opportunistic networking paradigms are considered to have broad potential applications, so far very little is known about which nodes are more important in both sustaining the network topology and forwarding or disseminating messages. To address this issue, this paper redefines the concept of walk and extends traditional Katz Centrality measurement to dynamic opportunistic social networks. Based on the Time Evolving Graph model, we derive a convenient formula to identify each node's information dissemination capability through computing the product of the adjacent matrix of each snapshot along the direction of time. The resulting matrix, in which the spatial and temporal dependency of the network nodes are fully captured, can conveniently be used to evaluate each node's relative dissemination capability. We apply our method to two real experiment trace datasets and the results show that, several mobile nodes with highest communicability identified by our method are more efficient in message dissemination than the others in the whole network. Those nodes can be chosen as good candidates when some interventions, such as accelerating or suppressing the speed of information spreading in network, are required to be made on network.

A tree-based topology is often used to flood packets from the sink node in wireless sensor networks (WSNs). Therefore, flooding tree construction is an important and fundamental problem in WSNs, and has been extensively investigated in the literature. However, we note that the flooding tree construction problem in asynchronous low-duty-cycle WSNs has not been sufficiently investigated in existing work. In this work, we focus our investigation on minimum-delay and energy-efficient flooding tree construction considering the duty-cycle operation and unreliable wireless links. We formulate the problem as a undetermined-delay-constrained minimum spanning tree (UDC-MST) problem, where the delay constraint is known a posteriori. We design a distributed heuristic algorithm, named MDET, to solve the problem. Through extensive simulations, we demonstrate that MDET achieves a very good balance between flooding delay and energy efficiency.

Indoor localization is essential to enable location-based services in wireless pervasive computing environment. In recent years, WiFi fingerprint-based localization has received considerable attention due to its deployment practicability. In order to achieve on-the-fly localization, WiFi receivers (e.g., mobile phones or laptops) being located need to scan WiFi signals continuously. Since they are normally battery driven, energy efficiency is a very important consideration in WiFi fingerprinting localization systems. Motivated by the fact that IEEE 802.11 (WiFi) and 802.15.4 (ZigBee) channels overlap in the 2.4GHz ISM band, in this work, we develop a WiFi fingerprint-based localization system using ZigBee radio, called ZiLoc. We first present a novel RSS-location fingerprint model to identify the features of surrounding APs. We then propose a simple yet effective method to compute the similarity of two RSS fingerprints. Experimental results demonstrate that ZiLoc can achieve an average of 85% room-level localization accuracy and reduce more than 60% energy consumption compared with the method using WiFi interfaces to collect RSS fingerprints.

Mobile devices with local wireless interfaces can be organized into opportunistic networks which exploit communication opportunities arising from the movement of their users. With the proliferation and increasing capabilities of these devices, it is significant to investigate message dissemination over opportunistic networks to maximize the potential of those networks. In this paper, we analyze the performance of copy-limited flooding over opportunistic networks, where each node can only send no more than k copies of the same message. For this purpose, we propose a network model called Markov and Random graph Hierarchic Model (MRHM), where a node transfers among different Main-areas (places frequently visited by nodes) according to the Markov rule, and two different nodes in the same Main-area can establish a connection with a certain probability. We theoretically analyze the performance of k-copy limited flooding over MRHM in terms of delivery rate and delay. Our extensive experiments over MRHM and real traces reveal that when k equals 3, the performance of k-copy limited flooding is very close to that of Epidemic Routing.

Chip multiprocessor (CMP) techniques have been implemented in embedded systems due to tremendous computation requirements. Three-dimension (3D) CMP architecture has been studied recently for integrating more functionalities and providing higher performance. The high temperature on chip is a critical issue for the 3D architecture. In this article, we propose an online thermal prediction model for 3D chips. Using this model, we propose novel task scheduling algorithms based on rotation scheduling to reduce the peak temperature on chip. We consider data dependencies, especially inter-iteration dependencies that are not well considered in most of the current thermal-aware task scheduling algorithms. Our simulation results show that our algorithms can efficiently reduce the peak temperature up to 8.1 ˆ C.

The majority of previous studies have focused on the analyses of an entire graph (network) or the giant connected component in a graph. Here we study the disconnected components (non-giant connected components) in real social networks, and reporting some interesting discoveries on how these disconnected components evolve over time. We study six diverse, real networks (citation networks, online social networks, academic collaboration networks, and others), and make the following major contributions: (a) we make empirical observations of the longevity distribution of disconnected components, and find that the curve of the distribution demonstrates a decaying trend; (b) we find that the distributions of final size of disconnected components that merge with one another or get absorbed by the giant connected component both follow power laws; (c) we find that the majority of mergings are between disconnected components and the giant connected component. The mergings that happen among disconnected components are small in scale (involve only a few components). The longevity distributions of the disconnected components in those mergings are similar, where the shortest-lived disconnected components are the most in number; and (d) we propose an empirical generative model that can produce the networks with our observed patterns.

This paper proposes a clustering method, group-driven process in the sociology of the clustering process simulation, whose clustering process simulates the driven process of colony formation in sociology. Clustering experiments show that the clustering accuracy and clustering speed of the algorithm in complex network are superior to the classic optimization Fast-Newman clustering algorithm.

Reliable flooding in wireless sensor networks (WSNs) is desirable for a broad range of applications and network operations. However, it is a challenging problem to ensure 100% flooding coverage efficiently considering the combined effects of low-duty-cycle operation and unreliable wireless transmission. In this work, we propose a novel dynamic switching-based reliable flooding (DSRF) framework, which is designed as an enhancement layer to provide efficient and reliable flooding over a variety of existing flooding tree structures in low-duty-cycle WSNs. Through comprehensive simulations, we demonstrate that DSRF can effectively improve both flooding energy efficiency and latency.

One fundamental requirement of the trust-aware recommender system (TARS) is to efficiently find as many recommenders as possible for the active users. Existing approaches of TARS choose to search the entire trust network, which have very high computational cost. Though the trust network is the scale-free network, we show via experiments that TARS cannot find satisfactory number of recommenders by directly applying the classical searching mechanism of the scale-free network. This is because it chooses the local highest-degree node at each step of the trust propagation. Since the power of the trust network's degree distribution is not big enough, the selected nodes cannot cover superior number of users. In this paper, we propose an efficient searching mechanism, named S_Searching, for TARS based on the scale-freeness of trust networks: choosing the global highest-degree nodes to construct a Skeleton, and searching the recommenders via this Skeleton. Benefiting from the superior outdegrees of the nodes in the Skeleton, S_Searching can find the recommenders very efficiently. Experimental results show that S_Searching can find almost the same number of recommenders as that of conducting full search, which is much more than that of applying the classical searching mechanism in the scale-free network, while the computational complexity and cost is much less.

Recently, research on complex network theory and applications draws a lot of attention in both academy and industry. In mobile ad hoc networks (MANETs) area of research, a critical issue is to design the most effective topology for given problems. It is natural and significant to consider complex networks topology when optimising the MANET topology. Current works usually transform MANET or sensor network topologies into either small-world or scale-free. However, some fundamental problems remain unsolved. Specifically, what are the average shortest path length, degree distribution and clustering characteristics of MANETs? Do MANETs have small-world effect and scale-free property? In this work, the authors introduce complex networks theory into the context of MANET topology and study complex network properties of the MANETs to answer the above questions. The authors have theoretically analysed the degree distribution and clustering coefficient of MANETs and proposed approach to computing them. The degree distribution and clustering coefficient of MANETs are theoretically deduced from node space probability distribution on different mobility models (including but not limited to random waypoint model). Simulation results on average shortest path length, clustering coefficient and degree distribution show that in most cases MANETs do not have the small-world effect and scale-free property.

A distributed mobile system consists of a group of heterogeneous mobile devices connected by wireless network. Due to the fact that most of the mobile devices are battery based, the lifetime of a mobile system depends on both the battery behavior and the energy consumption characteristics of tasks. In this paper, we present a set of models for task scheduling in mobile systems equipped with Dynamic Voltage Scaling (DVS) processors. We propose battery-aware algorithms to obtain task schedules satisfying the battery lifetime constraints. The simulations with randomly generated Directed Acyclic Graphs (DAG) show that our proposed algorithms generate better schedules that can satisfy the battery lifetime constraints.

A VANET consists of vehicles equipped with on board units (OBUs) that can communicate with each other and the road side base stations. Due to the mobility and sparse distribution of vehicles, the delivery delay of messages in the VANET is mainly caused by the message transmissions between vehicles. The message delivery delay directly impacts the deployment of applications in VANET, and hence, an in-depth study of message delivery delay in the VANET is significant. In this paper, we focused on the investigation of the message delivery delay in the V2V stage with a bidirectional setting. The bidirectional traffic was modeled as a combination of two Poisson point processes. Based on the sub-additive ergodic theory, we found theoretically that the message delivery delay has a linear relationship with the message forwarding distance. Further, the upper bound of the coefficient of the linear relationship has an exponential polynomial relation with the density of vehicles on the road and decreases with the increment of the velocity of the traffic.

With their proliferation and increasing capabilities, mobile devices with local wireless interfaces can be organized into opportunistic networks that exploit communication opportunities arising out of the movement of their users. Because the nodes are carried by people, these opportunistic networks can also be viewed as social networks. Unfortunately, existing routing algorithms for opportunistic networks rely on relatively simple mobility models that rarely consider these social network characteristics. In this paper, we propose PreS (Predict and Spread), an efficient routing algorithm for opportunistic networking that employs an adapted Markov chain to model a node's mobility pattern, and capture its social characteristics. A comparison with state-of-the-art algorithms suggests that PreS can yield better performance in terms of delivery ratio and delivery latency, and approaches the performance of the Epidemic algorithm with lower resource consumption.

Chip multiprocessor technique has been implemented in embedded systems due to the tremendous computation requirements. Three dimension chip multiprocessor architecture has been studied recently for integrating more functionalities and providing higher performance. The high temperature on chip is a critical issue for the 3D architecture. In this paper, we propose an online thermal prediction model for 3D chip. Using this model, we present a task scheduling algorithm based on rotation scheduling to reduce the peak temperature on chip. We consider the data dependencies, especially the inter-iteration dependencies which are not well considered in most of the current thermal-aware task scheduling algorithms. Our simulation result shows that our algorithm can efficiently reduce the peak temperature up to 10°C.

In cloud computing, computational resources are provided to remote users in the form of leases. For a cloud user, he/she can request multiple cloud services simultaneously. In this case, parallel processing in the cloud system can improve the performance. When applying parallel processing in cloud computing, it is necessary to implement a mechanism to allocate resource and schedule the tasks execution order. Furthermore, a resource allocation mechanism with preemptable task execution can increase the utilization of clouds. In this paper, we propose an adaptive resource allocation algorithm for the cloud system with preemptable tasks. Our algorithms adjust the resource allocation adaptively based on the updated of the actual task executions. And the experimental results show that our algorithms works significantly in the situation where resource contention is fierce.

In this paper we present Stroke++, a novel hybrid Chinese input method for touch screen mobile phones that leverages hieroglyphic properties of Chinese characters to enable faster and easier input of Chinese characters on mobile phones. By using a special keypad layout, a friendly user interface and an adaptive radical selection algorithm, we achieved a competitive inputting performance compared with currently prevalent mobile Chinese input methods, while keeping a low entry barrier for Chineseinput novices. An extensive evaluation results show that Stroke++ out-performs the state-of-the-art keystroke-based or handwriting recognition-based Chinese character inputting methods, as far as the input speed and convenience are concerned.

An infrastructure-as-a-service cloud system provides computational capacities to remote users. Parallel processing in the cloud system can shorten the execution of jobs. Parallel processing requires a mechanism to scheduling the executions order as well as resource allocation. Furthermore, a preemptable scheduling mechanism can improve the utilization of resources in clouds. In this paper, we present a preemptable job scheduling mechanism in cloud system. We propose two feedback dynamic scheduling algorithms for this scheduling mechanism. We compare these two scheduling algorithms in simulations. The results show that the feedback procedure in our algorithms works well in the situation where resource contentions are fierce.

The P2P streaming system must have the adaptation ability to the heterogeneity of the peer nodes and networks. In this paper, an adaptive scheme on video streaming over multi-hop path in P2P networks, called server-aided intermediate node adaptive transmission (SA-INAT), is proposed. In the scheme, the streaming server probes the network parameters in the multi-hop path periodically, and adjusts the output bit-rate and the encoding structure according to the probing result. Before relay the obtained streaming data, the intermediate peer nodes elegantly discard some frames to adapt the output rate to the bandwidth while alleviating the degradation in the decoding video quality in a delay-constrained manner. The experimental results show that the video data can be streamed over the multi-hop path efficiently and adaptively by the cooperation among the streaming server and the intermediate nodes.

In a wireless sensor network for data-gathering applications, if all network data congregate to a stationary sink node hop by hop, the sensor nodes near the sink have to consume more energy on forwarding data for other nodes, which probably causes the early function loss of the sensor network. Employing a mobile sink can alleviate the hotspot problem and balance the energy consumption among the sensor nodes. In this paper, we propose two autonomous moving schemes for the mobile sink. In our schemes, the sink makes moving decisions without complete knowledge of network topology and the energy distribution of all sensor nodes. We evaluated the performance of our moving schemes by simulation and the results show that both the two schemes can extend the network lifetime prominently.

In this paper, the influence of wireless channel fading on packet loss and delay for wireless local area networks is studied in detail. It shows that in CSMA/CA MAC scheme, packets with bad channel state will engross MAC buffer and deteriorate other downlink flows' performance. To overcome this problem, a novel channel state aware scheduling algorithm is presented to reduce the QoS degradation caused by channel fluctuation. It schedules every downlink flow into separated LLC layer queue firstly. Then, aided by the physical layer information, SNR, a channel flag based queue selection algorithm is used to isolate traffics with high PER (packet error rate). Simulation study shows this cross-layer scheme can decrease sending delay and packet loss for downlink real-time flows, and use wireless link resource more effectively