Zhaokun Wang, Jinyu Guo, Jingwen Pu, Hongli Pu, Meng Yang, Xunlei Chen, Jie Ou, Wenyi Li, Guangchun Luo, Wenhong Tian. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Modern industrial environments generate massive streams of discrete textual logs that record device status, error codes, and operational events. Deploying models on resource-constrained edge devices demands extreme data compression and fast inference, yet existing dataset distillation (DD) methods are designed for images and capture only short-term training dynamics via single-step gradient or distribution matching. To address these limitations, we propose textual trajectory matching (TTM), a novel textual DD framework that aligns student trajectories with expert trajectories derived from industrial data training. We introduce a Mask-and-Fill initialization to enhance trajectory diversity, expanding semantic representations. We further propose a manifold distribution-based initialization to preserve original semantic features through low-dimensional manifold analysis. To address model computational costs, we design a subset matching strategy to reducing by aligning critical structural components. Evaluated on SST-2, MNLI-m, AGNews, and BlueGene/L (BGL) with$\text{BERT}_{\text{BASE}}$,$\text{RoBERTa}_{\text{BASE}}$,$\text{XLNet}_{\text{BASE}}$, and LLaMA 3, TTM achieves 3.6% higher accuracy than state-of-the-art methods, enabling efficient industrial dataset and model deployment in smart factories.

Knowledge distillation has been proven to be an effective method for enhancing model performance, particularly in the domain of model compression. In this study, we propose a comprehensive approach that utilizes Brownian Distance Covariance (BDC) to measure the discrepancy between the logits produced by the teacher and student models. Unlike the conventional KL divergence used in traditional knowledge distillation, BDC captures not only linear relationships but also nonlinear dependencies, thereby overcoming the limitations of KL divergence and enabling the student model to learn more effectively from the teacher model. Additionally, our method aligns the discrepancies between the teacher and student models from both intra-class and inter-class perspectives. Extensive experiments demonstrate that our method achieves state-of-the-art (SOTA) performance across various network architectures and datasets. The code and resources related to this work are available at the following link: https://github.com/hengyin23654/BDCKD.

Post-training quantization (PTQ) has emerged as a promising approach for converting full-precision models into compact, low-precision models with minimal computational overhead, making them ideal for deployment in resource-constrained edge scenarios. While most existing PTQ techniques focus on minimizing the numerical discrepancy between model activations before and after quantization, such methods often overlook the inherent noise and distributional shifts caused by quantization, which can lead to severe performance degradation. To address this, we propose Distribution-Enhanced Reconstruction for PTQ (DEQuant), a novel approach that enhances the performance of quantized models by introducing a module that further enhances the alignment of activation pre- and post-quantization during model reconstruction. Extensive experiments demonstrate the effectiveness of DEQuant in several low-bit settings, achieving superior performance compared to existing methods. For instance, DEQuant achieves 14.18% accuracy on MobileNetV2 under the W2A2 configuration, representing a 5.72% improvement over the baseline QDrop and surpassing other baselines by 1–3%.

Deep neural networks for SAR image recognition often require compression for deployment on remote sensing platforms with limited computational and storage resources. Knowledge distillation (KD) is a key approach to improving the accuracy of lightweight networks. However, existing KD methods face challenges when applied to SAR images due to the small dataset size and the high noise in SAR images. To address this, this paper proposes a novel knowledge distillation method that relaxes the requirement for a strict linear relationship between the outputs of lightweight and large models, focusing instead on maintaining a Monotonic Relationship (MRKD). This reduces the difficulty of the KD task. Experiments on various SAR image classification and object detection datasets demonstrate that MRKD achieves state-of-the-art performance improvements for lightweight networks.

Exploration in sparse reward environments is a long-standing challenge in reinforcement learning. While advanced methods often enhance exploration by introducing intrinsic rewards, they may lead agents to local optima in procedurally-generated environments with high uncertainty and diversity. Specifically, agents may over rely on certain intrinsic signals, neglecting long-term tasks tied to external rewards. To address this, we leverage multimodal causal reasoning to identify key causal variables relevant to external tasks and propose a Causal Reasoning-guided intrinsic Goal generation method (CRG). Our approach integrates multimodal data to abstract invariant causal relationships in dynamic environments, using a teacher network to evaluate and select intrinsic goals with the strongest causal impact on task completion. This enables agents to explore and learn efficiently, enhancing overall task performance. We evaluated our method in several MiniGrid environments of varying complexity. Experimental results show that CRG can leverage causal relationships to generate more effective internal goals, significantly accelerating task completion and improving exploration efficiency.

While large language models (LLMs) show considerable promise across various fields, they have notable limitations in handling multi-document question answering (Multi-doc QA) tasks. The first challenge is long-range dependency modeling, where LLMs struggle to focus on key information in long texts, which weakens important semantic connections. Second, most LLMs suffer from the ''lost-in-the-middle'' issue, where they have difficulty processing information in the middle of long inputs. Current solutions either truncate global dependencies or demand costly finetuning, ultimately lacking a universal and simple solution for these challenges. To resolve these limitations, we propose Dual-Stage Adaptive Sharpening (DSAS) containing two modules. (i) The Contextual Gate Weighting (CGW) module alleviates ''lost-in-the-middle'' by assessing paragraph relevance through layer-wise attention tracking and position-aware weighting. (ii) The Reciprocal Attention Suppression (RAS) module enhances focus on critical paragraphs by suppressing information exchange between key and irrelevant texts, thus mitigating the limitations in long-range dependency modeling. Extensive experiments on four benchmarks demonstrate DSAS's efficacy across mainstream LLMs (Llama, Qwen, Mistral, and Deepseek), with an average F1-score improvement of 4.2% in Multi-doc QA tasks on Llama-3.1-8B-Instruct and Qwen2.5-14B-Instruct. Ablation studies confirm the essential contributions of both the CGW and RAS modules. In addition, detailed discussions in the Appendix further validate the robustness and scalability of DSAS.

With the continuous advancement of big data and artificial intelligence technologies, event prediction is increasingly being utilized across a multitude of domains. Predicting events allows for the exploration of the developmental trajectories and summarization of patterns associated with these events. However, events typically encompass a myriad of elements and intricate relationships, necessitating an enhancement in the precision of event prediction. However, the existing methods suffer from poor data quality, insufficient feature information, limited generalization capability of the models, and difficulties in evaluating prediction errors. This paper proposes a novel event prediction method based on graph representation learning, aiming to improve the accuracy of event prediction while reducing the time cost. By constructing causal graphs and introducing the script event simulation method, the architecture combines graph neural networks (GNNs) with BERT to simplify the event prediction process. Additionally, by combining GNNs with pretrained language models, a dynamic graph representation learning method is proposed. This means that a unified graph representation learning model can be built by following specific rules, thus predicting the development trajectory of events more accurately. The study evaluates the effectiveness of dynamic graph representation learning technology in a specific scenario, specifically in the context of employee career choices. By converting the career graph of employees into low‐dimensional representations, the effectiveness of the dynamic graph representation learning method in predicting employee career decisions is validated. This innovation not only improves the accuracy of event prediction but also helps better understand and respond to complex event relationships in practical applications, providing decision‐makers with more powerful information support. Therefore, this research has important theoretical and practical significance, providing valuable references for future studies in related fields.

Recent advancements in language models have greatly improved dialogue systems, but they still face challenges in generating personalized responses that are consistent with the user’s persona and dialogue context. Existing approaches typically model dialogue context and persona information together in a unified manner, but they lack fine-grained differentiation between the two, leading to inaccurate user modeling. This misalignment hinders the ability of dialogue systems to produce personalized responses. In this work, we propose cross-graph knowledge exchange (CKE), a novel algorithm designed to enhance personalized response generation in dialogue systems. CKE constructs separate dialogue user graphs for each party in the dialogue, representing both their dialogue context and persona information. These graphs are then utilized to perform cross-graph structured knowledge aggregation, where the aggregation is under supervised by both its own and the other party’s persona and dialogue context, providing richer, more accurate representations. Furthermore, CKE introduces a hybrid prompt template that combines both discrete and continuous elements, improving the language model’s ability to leverage graph-structured information. The experimental results demonstrate that CKE significantly outperforms existing baseline methods in generating more coherent, contextually appropriate, and personalized responses in dialogue systems.

Large language models (LLMs) have gained a lot of attention and achievements recently because of their significant comprehension and generative abilities. However, the large-scale parameters of LLMs require considerable computational resources in the training and inference process, which restricts their wide application. To overcome this challenge, we propose an efficient mixed precision weight quantization (EMWQ) method for LLMs in this article. Specifically, we introduce a new outlier detection method by analyzing the weight distribution instead of the conventional weight magnitude. Then, we propose a dual-quantization strategy that quantizes both the outlier critical columns and the residual matrices with different precision. Besides, we introduce two effective EMWQ-based application frameworks, the EMWQ-R and EMWQ-O in our study. Comprehensive experiments are conducted on the Penn Treebank (PTB), C4, ARC-Easy datasets, and MMLU benchmark across various tasks. The comparison results demonstrate that the proposed EMWQ achieves state-of-the-art performance in mixed precision quantization and further reduces computational memory cost. Besides, it has higher generalizability compared with conventional methods.

Traditional spiking learning algorithm aims to train neurons to spike at a specific time or on a particular frequency, which requires precise time and frequency labels in the training process. While in reality, usually only aggregated labels of sequential patterns are provided. The aggregate-label (AL) learning is proposed to discover these predictive features in distracting background streams only by aggregated spikes. It has achieved much success recently, but it is still computationally intensive and has limited use in deep networks. To address these issues, we propose an event-driven spiking aggregate learning algorithm (SALA) in this article. Specifically, to reduce the computational complexity, we improve the conventional spike-threshold-surface (STS) calculation in AL learning by analytical calculating voltage peak values in spiking neurons. Then we derive the algorithm to multilayers by event-driven strategy using aggregated spikes. We conduct comprehensive experiments on various tasks including temporal clue recognition, segmented and continuous speech recognition, and neuromorphic image classification. The experimental results demonstrate that the new STS method improves the efficiency of AL learning significantly, and the proposed algorithm outperforms the conventional spiking algorithm in various temporal clue recognition tasks.

Attempting to fully exploit the rich information of topological structure and node features for attributed graph, we introduce self-supervised learning mechanism to graph representation learning and propose a novel Self-supervised Consensus Representation Learning (SCRL) framework. In contrast to most existing works that only explore one graph, our proposed SCRL method treats graph from two perspectives: topology graph and feature graph. We argue that their embeddings should share some common information, which could serve as a supervisory signal. Specifically, we construct the feature graph of node features via k-nearest neighbour algorithm. Then graph convolutional network (GCN) encoders extract features from two graphs respectively. Self-supervised loss is designed to maximize the agreement of the embeddings of the same node in the topology graph and the feature graph. Extensive experiments on real citation networks and social networks demonstrate the superiority of our proposed SCRL over the state-of-the-art methods on semi-supervised node classification task. Meanwhile, compared with its main competitors, SCRL is rather efficient.

Recently, the surge in popularity of the Internet of Things (IoT), mobile devices, social media, etc., has opened up a large source for graph data. Graph embedding has been proved extremely useful to learn low-dimensional feature representations from graph-structured data. These feature representations can be used for a variety of prediction tasks from node classification to link prediction. However, the existing graph embedding methods do not consider users' privacy to prevent inference attacks. That is, adversaries can infer users' sensitive information by analyzing node representations learned from graph embedding algorithms. In this article, we propose adversarial privacy graph embedding (APGE), a graph adversarial training framework that integrates the disentangling and purging mechanisms to remove users' private information from learned node representations. The proposed method preserves the structural information and utility attributes of a graph while concealing users' private attributes from inference attacks. Extensive experiments on real-world graph data sets demonstrate the superior performance of APGE compared to the state-of-the-arts. Our source code can be found at https://github.com/KaiyangLi1992/Privacy-Preserving-Social-Network-Embedding.

Significant progress has been made in recent years in image captioning, an active topic in the fields of vision and language. However, existing methods tend to yield overly general captions and consist of some of the most frequent words/phrases, resulting in inaccurate and indistinguishable descriptions (see Fig. 1). This is primarily due to (i) the conservative characteristic of traditional training objectives that drives the model to generate correct but hardly discriminative captions for similar images and (ii) the uneven word distribution of the ground-truth captions, which encourages generating highly frequent words/phrases while suppressing the less frequent but more concrete ones. In this work, we propose a novel global-local discriminative objective that is formulated on top of a reference model to facilitate generating fine-grained descriptive captions. Specifically, from a global perspective, we design a novel global discriminative constraint that pulls the generated sentence to better discern the corresponding image from all others in the entire dataset. From the local perspective, a local discriminative constraint is proposed to increase attention such that it emphasizes the less frequent but more concrete words/phrases, thus facilitating the generation of captions that better describe the visual details of the given images. We evaluate the proposed method on the widely used MS-COCO dataset, where it outperforms the baseline methods by a sizable margin and achieves competitive performance over existing leading approaches. We also conduct self-retrieval experiments to demonstrate the discriminability of the proposed method.

Visual relationship can capture essential information for images, like the interactions between pairs of objects. Such relationships have become one prominent component of knowledge within sparse image data collected by multimedia sensing devices. Both the latent information and potential privacy can be included in the relationships. However, due to the high combinatorial complexity in modeling all potential relation triplets, previous studies on visual relationship detection have used the mixed visual and semantic features separately for each object, which is incapable for sparse data in IoT systems. Therefore, this paper proposes a new deep learning model for visual relationship detection, which is a novel attempt for cooperating computational intelligence (CI) methods with IoTs. The model imports the knowledge graph and adopts features for both entities and connections among them as extra information. It maps the visual features extracted from images into the knowledge‐based embedding vector space, so as to benefit from information in the background knowledge domain and alleviate the impacts of data sparsity. This is the first time that visual features are projected and combined with prior knowledge for visual relationship detection. Moreover, the complexity of the network is reduced by avoiding the learning of redundant features from images. Finally, we show the superiority of our model by evaluating on two datasets.

The huge amount of graph data are published and shared for research and business purposes, which brings great benefit for our society. However, user privacy is badly undermined even though user identity can be anonymized. Graph de-anonymization to identify nodes from an anonymized graph is widely adopted to evaluate users' privacy risks. Most existing de-anonymization methods which are heavily reliant on side information (e.g., seeds, user profiles, community labels) are unrealistic due to the difficulty of collecting this side information. A few graph de-anonymization methods only using structural information, called seed-free methods, have been proposed recently, which mainly take advantage of the local and manual features of nodes while overlooking the global structural information of the graph for de-anonymization.

In Network Function Virtualization (NFV), to satisfy the Service Functions (SFs) requested by a customer, service providers will composite a Service Function Chain (SFC) and embed it onto the shared Substrate Network (SN). For many latency-sensitive and computing-intensive applications, the customer forwards data to the cloud/server and the cloud/server sends the results/models back, which may require different SFs to handle the forward and backward traffic. The SFC that requires different SFs in the forward and backward directions is referred to as hybrid SFC (h-SFC). In this paper, we, for the first time, comprehensively study how to optimize the latency in Hybrid SFC composition and Embedding (HSFCE). When each substrate node provides only one unique SF, we prove the NP-hardness of HSFCE and propose the first 2-approximation algorithm to jointly optimize the processes of h-SFC construction and embedding, which is called Eulerian Circuit based Hybrid SFP optimization (EC-HSFP). When a substrate node provides various SFs, we extend EC-HSFP and propose the efficient Betweenness Centrality based Hybrid SFP optimization (BC-HSFP) algorithm. Our extensive simulations and analysis show that EC-HSFP can hold the 2-approximation, while BC-HSFP outperforms the algorithms directly extended from the state-of-art techniques by an average of 20%.

Nowadays, the social network becomes an indispensable part of people's daily life, meanwhile offers an unprecedentedly convenient access for purposive individuals to influence the opinions of network users. Current studies present a subtle persuasion approach that finds a number of key users meanwhile varies their susceptibility extent to impact the public opinion. Such persuasion is significantly critical for public security, as it could facilitate both the spreading and dispelling of malicious rumors. However, the major body of these studies enclose impractical assumptions, such that persuaders have an unlimited budget, or the costs of varying different users' susceptibilities are the same, thus rendering these works unsuitable for realistic scenarios. Therefore, this work originally proposes a more practical and generalized problem of persuasion, where varying the susceptibilities of different users holds different costs. The analysis of its non-convexity, non-submodularity and complexity shows that solving the proposed problem is nontrivial, thus inspiring us to provide an intuitive greedy algorithm. Furthermore, we design an accelerated algorithm based on the community property, which reduces the time consumption more than one order of magnitude. The acceleration is based on the intuition that the impact of a user within a proper community could be a good estimation of the impact in the whole network, while the computation of the former one is much more efficient. The relationship between two algorithms is fully analyzed, which shows the community-based algorithm can degenerate to the intuitive greedy algorithm under a specific setting. Finally, comprehensive evaluations on real-world datasets show the superiority of proposed algorithms on both effectiveness and efficiency.

Finding a suitable data representation for a specific task has been shown to be crucial in many applications. The success of subspace clustering depends on the assumption that the data can be separated into different subspaces. However, this simple assumption does not always hold since the raw data might not be separable into subspaces. To recover the "clustering-friendly" representation and facilitate the subsequent clustering, we propose a graph filtering approach by which a smooth representation is achieved. Specifically, it injects graph similarity into data features by applying a low-pass filter to extract useful data representations for clustering. Extensive experiments on image and document clustering datasets demonstrate that our method improves upon state-of-the-art subspace clustering techniques. Especially, its comparable performance with deep learning methods emphasizes the effectiveness of the simple graph filtering scheme for many real-world applications. An ablation study shows that graph filtering can remove noise, preserve structure in the image, and increase the separability of classes.

The collection of high-confidence data has been one prominent step for many services in smart city systems. However, the privacy issues have been thwarting the seamless publication of data, especially as the data from different aspects of daily life may provide unprecedented coverage of contributors. Current solutions have been carefully designed to perturb or suppress the data before publication, so as to balance the privacy and utilities. However, they cannot fit the practice in smart cities, where multiple service providers request information on heterogeneous domains and regions of the city. Therefore, this article proposes a novel framework for data publication of workers in smart city systems. The framework allows workers and requestors to own and request various types of contents in different regions. The objective is to maximize the number of service providers receiving qualified utilities under privacy constraints. Furthermore, differential privacy is applied to guarantee that workers will not disclose personal information to requestors. In the technical part, the problem is proved to be NP-complete. Then two algorithms and strategies are proposed toward different cases: 1) workers apply identical privacy budgets for all published data and 2) workers are flexible on privacy settings. Both algorithms are theoretically analyzed on their performance of the released results. Finally, the evaluation of data sets of local businesses reveals that proposed algorithms can outperform baseline methods.

The rapid growth of DNA-sequencing technologies motivates more personalized and predictive genetic-oriented services, which further attract individuals to increasingly release their genome information to learn about personalized medicines, disease predispositions, genetic compatibilities, etc. Individual genome information is notoriously privacy-sensitive and highly associated with relatives. In this paper, we present an inference attack algorithm to predict target genotypes and phenotypes based on belief propagation in factor graphs. With this algorithm, an attacker can effectively predict the target genotypes and phenotypes of target individuals based on genome information shared by individuals or their relatives, and genotype and phenotype association from genome-wide association study (GWAS). To address the privacy threats resulted from such inference attacks, we elaborate the metrics to evaluate data utility and privacy and then present a data sanitization method. We evaluate our inference attack algorithm and data sanitization method on real GWAS dataset: Age-related macular degeneration (AMD) case/control dataset. The evaluation results show that our work can effectively defense against genome threats while guaranteeing data utility.

Person re-identification is the task of recognizing an individual across heterogeneous non-overlapping camera views. It has become a crucial capability needed by many applications in public space video surveillance. However, it remains a challenging task due to the subtle inter-class similarity and large intra-class variation found in person images. Current CNN-based approaches have focused and investigated traditional identification or verification frameworks. Such approaches typically use the whole input image including the background and fail to pay attention to specific body parts, deviating the feature representation learning from informative parts. In this article, we introduce a self-attention mechanism coupled with cross-resolution to improve the feature representation learning of person re-identification task. The proposed self-attention module reinforces the most informative parts from a high-resolution image using its internal representation at the low-resolution. In particular, the model is fed with a pair of images on a different scale and consists of two branches. The upper branch processes the high-resolution image and learns high dimensional feature representation while the lower branch processes the low-resolution image and learns a filtering attention heatmap. The feature maps on the lower branch are subsequently weighted to reflect the importance of each patch of the input image using a softmax operation; whereas, on the upper branch, we apply a max pooling operation to downsample the high-resolution feature map before element-wise multiplied with the attention heatmap. Our attention module helps the network learn the most discriminative visual features of multiple regions of the image and is specifically optimized to attend and enforce feature representation at different scales. Extensive experiments on three large-scale datasets show that network architectures augmented with our self-attention module systematically improve their accuracy and outperform various state-of-the-art models by a large margin.

With the technology of Network Function Virtualization (NFV), a multicast service (e.g., real-time multimedia streaming or event monitoring) may be accommodated with a Service Function Chain (SFC). The SFC consists of an ordered set of network functions running on generic physical hardware to provide services from source node to each destination node of the multicast service. In this paper, we define the problem of Multicast-aware Service Function Tree Embedding (M-SFTE), which allows the multicast flows to traverse through network service functions before reaching destination nodes. The M-SFTE maps user's SFC-based multicast requests onto a shared substrate network while considering the constraints of network functionality, computing demand of each virtual network function node, and the bandwidth demand of the request. We propose a novel algorithm, called Minimum Cost Multicast Service Function Tree (MC-MSFT) to jointly optimize the process of constructing SFC-based multicast tree and allocating requested resource to embed the tree onto the physical network. The experimental results show that the MC-MSFT algorithm outperforms the traditional greedy-based algorithms as much as by 35% in terms of the total bandwidth consumption.

Network Function Virtualization (NFV) takes advantages of the emerging technologies in virtualization and automation to offer new ways in design, deployment, and management of networking services. In NFV, the proprietary hardware-based network functions are replaced by the software-based modules named as Virtual Network Functions (VNFs) or Service Functions (SFs). A network service request from the customer can be formed by multiple SFs. To satisfy a network service request, the service provider has to chain the SFs in the request into a Service Function Chain (SFC) and embed the constructed SFC onto the shared substrate network. In this paper, we comprehensively study how to composite and embed an SFC onto a shared substrate network with unique service function. We formulate this problem with the Integer Linear Programming (ILP) technique. We also propose an efficient heuristic algorithm with 2-approximation boundary, namely, Spanning Closed Walk based SFC Embedding (SCW-SFCE). Our extensive simulations and analysis show that the proposed approach can achieve near-optimal performance in a small network and outperform the Nearest Neighbour (NN) algorithm.

Network Function Virtualization (NFV) technology decouples network functions from proprietary hardware equipments. As a result, Internet Service Providers (ISPs) implement software-based network functions on generic highvolume substrate network devices. In NFV, a Service Function Chain (SFC) is defined as an ordered set of abstract network functions running on specific substrate nodes (e.g., servers). A challenging issue in NFV management and orchestration is how to optimize the Dependence-aware SFC Embedding in substrate Optical networks (D_SFCE_O). In this paper, we propose a novel algorithm, namely, Dependence-aware SFC embedding with Least-Used consecutive subcarriers (D_SFC_LU), which jointly optimizes SFC design, SFC mapping and spectrum allocation in optical networks. To minimize resource consumption, D_SFC_LU takes advantages of the proposed techniques: Impact Factor based Node Selection (IFNS), Chain Node Mapping (CNM) and Chain-Fit (CF) spectrum allocation. Our simulation and analysis demonstrate that D_SFC_LU can efficiently embed a network requests while minimizing the required substrate resource in optical networks.

In Network Function Virtualization (NFV), Service Function Chain (SFC) is composed of Virtual Network Function (VNF) nodes that are chained via VNF links. SFCs can be specified as unidirectional or bidirectional. A unidirectional SFC (u-SFC) demands the traffic being forwarded via the VNFs in one direction, while a bidirectional SFC (b-SFC) requires bidirectional traffic flows. In this paper, for the first time, we investigate the problem of how to efficiently deploy a hybrid SFC (h-SFC), whereas some VNF nodes are required to process bidirectional traffic while others only handle unidirectional traffic. We define a new problem called hybrid SFC Deployment (h-SFCD). When each substrate node provides one unique VNF, we prove the NP-hardness of the h-SFCD problem and propose an approximate algorithm, namely, 2-approximation Hybrid Service function chain Deployment in Unique function networks (2-HSD-U). Our experimental results show that the proposed 2-HSD-U algorithm significantly outperforms the heuristic algorithm based on the traditional Nearest-Neighbor technique.

With dramatic developments of blockchain technology, a number of blockchain-based applications emerge rapidly, among which the incorporation of blockchain into Internet of Things is one of the most valued research direction. Such powerful incorporation is a double-sided sword, i.e., it can benefit both individuals and society but has the vulnerability to coin hopping attack that is a new type of pool mining attack and hard to happen in traditional blockchain networks. In this paper, we theoretically prove the feasibility of coin hopping attack, deeply analyze the conditions of attack implementation, and comprehensively investigate the impacts of coin hopping attack. Moreover, some defense strategies are addressed. To our best knowledge, this paper is the first work targeting coin hopping attack.

In social networks, several influential individuals can promote an idea or a product to numerous individuals. Thus, it is valuable to solve the influence maximization (IM) problem, which asks for finding the most influential set of individuals in a social network. To estimate the influence of individuals, the existing independent cascade (IC) model simulates the influence diffusion only considering the influences from direct in-neighbors to nodes. This consideration does not hold in real life. In many cases, people are likely influenced by information depending on where it comes from, instead of who gives it. To simulate the influence diffusion more accurate, this paper proposes the second-order IC model, which takes the previous influence into consideration. In addition, we design an approximate algorithm and its distributed extension for IM under the second-order IC model. Experimental results show that our second-order IC model outperforms the IC model in terms of simulating influence diffusions. The proposed algorithms are efficient, and the obtained node sets are influential.

Transmission is crucial for big data analysis and learning in industrial Internet of Things (IoT) systems. To transmit data with limited energy is a challenge. This paper studies the problem of data transmission in energy harvesting systems with capacitor to supply energy where the energy receiving rate varies over time. The energy receiving rate is slower when the capacitor receives more energy. Based on this characteristic, we study the problem of how to transmit more data when the energy receiving time is not continuous. Given many packets that arrive at different time instances, there is a tradeoff between transmitting the packet right now or saving the energy to transmit the future arriving packets. We formalize two types of problems. The first one is how to minimize the total completion time when there is enough energy to transmit all the packets. The second one is how to transmit as many packets as possible when the energy is not enough to transmit all the packets. For the first problem, we give a 1 + α approximation off line algorithm when all the information of the packets and the energy receiving periods is known in advance, and a max{2, β} competitive ratio online algorithm where the information is not known in advance. For the second problem, we study three cases and give a 6 + [h/(b/R)] approximation off line algorithm for the general situation. We also prove that there does not exit a constant competitive ratio online algorithm.

The Internet of Things (IoT) has attracted the interests of both academia and industry and enables various real-world applications. The acquirement of large amounts of sensing data is a fundamental issue in IoT. An efficient way is obtaining sufficient data by the mobile crowdsensing. It is a promising paradigm which leverages the sensing capacity of portable mobile devices. The crowdsensing platform is the key entity who allocates tasks to participants in a mobile crowdsensing system. The strategy of task allocating is crucial for the crowdsensing platform, since it affects the data requester’s confidence, the participant’s confidence, and its own benefit. Traditional allocating algorithms regard the privacy preservation, which may lose the confidence of participants. In this paper, we propose a novel three-step algorithm which allocates tasks to participants with privacy consideration. It maximizes the benefit of the crowdsensing platform and meanwhile preserves the privacy of participants. Evaluation results on both benefit and privacy aspects show the effectiveness of our proposed algorithm.

Network virtualization in optical networks enables the decoupling of network services from the underlying hardware infrastructure to allow multiple Virtual Optical Requests (VORs) sharing the same Substrate/physical Optical Network (SON). The challenge of mapping VORs onto the shared SON lies on how to efficiently allocate physical resource for the VORs, which is referred to as Virtual Optical Network Embedding (VONE). Many recent research focus on the NP-Hard VONE optimization problem. In this paper, for the first time, we explore how to efficiently map a given VOR for a multicast service onto a shared SON while considering the fanout (splitting/forwarding) limitation of the physical optical switches. We propose a novel algorithm, namely, Centrality-based Degree Bounded Shortest Path Tree (C-DB-SPT) to minimize the resource usage while satisfying the degree limitation in the shared SON. The experimental results show that the C-DB-SPT algorithm outperforms the traditional greedy-based algorithms as much as by 35% in terms of the total bandwidth consumption.

In this paper, we propose a prototype-based classification model for evolving data streams, called SyncStream, which allows dynamically modeling time-changing concepts, making predictions in a local fashion. Instead of learning a single model on a fixed or adaptive sliding window of historical data or ensemble learning a set of weighted base classifiers, SyncStream captures evolving concepts by dynamically maintaining a set of prototypes in a proposed P-Tree, which are obtained based on the error-driven representativeness learning and synchronization-inspired constrained clustering. To identify abrupt concept drifts in data streams, PCA and statistical analysis based heuristic approaches have been introduced. To further learn the associations among distributed data streams, the extended P-Tree structure and KNN-style strategy are introduced. We demonstrate that our new data stream classification approach has several attractive benefits: (a) SyncStream is capable of dynamically modeling the evolving concepts from even a small set of prototypes. (b) Owing to synchronization-based constrained clustering and P-Tree, SyncStream supports efficient and effective data representation and maintenance. (c) SyncStream is also tolerant of inappropriate or noisy examples via error-driven representativeness learning. (d) SyncStream allows learning relationship among distributed data streams at the instance level. The experimental results indicate its efficiency and effectiveness.

Machine learning approaches have been widely used to tackle the problem of sensor array drift in E‐Nose systems. However, labeled data are rare in practice, which makes supervised learning methods hard to be applied. Meanwhile, current solutions require updating the analytical model in an offline manner, which hampers their uses for online scenarios. In this paper, we extended Target Domain Adaptation Extreme Learning Machine (DAELM_T) to achieve high accuracy with less labeled samples by proposing a Weighted Domain Transfer Extreme Learning Machine, which uses clustering information as prior knowledge to help select proper labeled samples and calculate sensitive matrix for weighted learning. Furthermore, we converted DAELM_T and the proposed method into their online learning versions under which scenario the labeled data are selected beforehand. Experimental results show that, for batch learning version, the proposed method uses around 20% less labeled samples while achieving approximately equivalent or better accuracy. As for the online versions, the methods maintain almost the same accuracies as their offline counterparts do, but the time cost remains around a constant value while that of offline versions grows with the number of samples.

With network virtualization over Elastic Optical Networks (EONs), network services are decoupled from the underlying hardware infrastructure to enable multiple Virtual Optical Requests (VORs) sharing the same Substrate/physical Optical Network (SON). The embedding process of VORs onto the shared SON while satisfying the computing resource and spectrum allocation constraints is referred to Virtual Optical Network Embedding (VONE), which is an NP-Hard problem. In this paper, for the first time, we investigate how to efficiently map a given VOR in the form of virtual optical multicast tree onto an SON. We propose a novel algorithm that is called Impact Factor based Virtual Optical Multicast Tree Embedding (IF-VOMTE) to minimize the resource usage and avoid redundant multicast transmission in the shared SON. The experimental results show that our algorithm outperforms the schemes based on traditional techniques such as Greedy Node Mapping (GNM-SP) and First-Fit Node Mapping (FFNM-SP) in terms of the total cost of bandwidth consumption and the reduction of redundant multicast transmission.

The emerging Network Function Virtualization (NFV) technology decouples network functions from the proprietary hardware, which allows the Internet Service Providers (ISPs) to implement network functions as software running on top of a physical (or substrate) node. With NFV, a Service Function Chain (SFC) is defined as an ordered set of network function instances running on specific substrate network nodes to provide services for client users. In this paper, we define the problem of Dependence- Aware Service Function Chain (D_SFC) design and mapping. We study how to efficiently accommodate user's D_SFC requests in the substrate network while considering the constraints of function dependence, computing demand of virtual nodes and bandwidth demand of the D_SFC. We propose a novel heuristic algorithm, called D_SFC design and resource allocation with Adaptive Mapping (D_SFC_AM), which jointly optimizes the processes of designing a D_SFC and allocating resources requested by the chain. D_SFC_AM employs the proposed techniques of dependence sorting and independent grouping that effectively take into account the node dependence and the resource status of the substrate network. Our experimental results show that the proposed algorithm significantly outperforms the scheme based on the traditional topological sorting in which the process of designing a D_SFC and allocating resources requested by the chain is done sequentially.

Network virtualization enables the decoupling of network services from the underlying hardware infrastructure to allow the same Substrate/physical Network (SN) shared by multiple Virtual Network (VN) requests. The process of mapping virtual nodes and links onto a shared SN while satisfying the computing and bandwidth constraints is referred to Virtual Network Embedding (VNE) as an NP-hard problem. In this paper, for the first time, we explore how to efficiently map a given Virtual Multicast Tree (VMT) request onto a substrate network. We propose a novel algorithm, namely, Virtual Multicast Tree Embedding based on dynamic Impact Factor (VMTE-IF) to minimize the required resource and redundant multicast transmission in the substrate network. The experimental results show that our algorithm outperforms the traditional greedy-based algorithms over 50% in terms of the cost of bandwidth consumption.

It is a daunting challenge to balance the convergence and diversity of an approximate Pareto front in a many-objective optimization evolutionary algorithm. A novel algorithm, named many-objective particle swarm optimization with the two-stage strategy and parallel cell coordinate system (PCCS), is proposed in this paper to improve the comprehensive performance in terms of the convergence and diversity. In the proposed two-stage strategy, the convergence and diversity are separately emphasized at different stages by a single-objective optimizer and a many-objective optimizer, respectively. A PCCS is exploited to manage the diversity, such as maintaining a diverse archive, identifying the dominance resistant solutions, and selecting the diversified solutions. In addition, a leader group is used for selecting the global best solutions to balance the exploitation and exploration of a population. The experimental results illustrate that the proposed algorithm outperforms six chosen state-of-the-art designs in terms of the inverted generational distance and hypervolume over the DTLZ test suite.

In network virtualization, the network services are decoupled from the underlying hardware infrastructure such that multiple virtual network requests can be mapped onto the same physical substrate network. The process of mapping virtual networks onto the substrate network with minimum resources while satisfying the constraints such as computing capacity, bandwidth and memory is referred to as virtual network embedding. In this paper, we investigate how to efficiently map a given virtual network with multicast services. We propose a closeness-centrality based multicast-aware virtual network embedding (CC-MVNE) algorithm to minimize the needed resources for the virtual nodes/links mapping and multicast transmission. Our extensive simulation and analysis show that the proposed approach outperforms the traditional greedy algorithm as much as by 40% in terms of transmission bandwidth consumption.

Network Function Virtualization (NFV) explores the virtualization technologies to offer Network-as-a- Service (NaaS) through connected virtual network functions. The network operations that were previously performed by specialized hardware are consolidated as software-based virtual network functions (VNFs). These VNFs can be implemented in the telecom clouds with high volume servers, switches and storage. With the NFV orchestration, a service function graph (SFG) can be built to provide network services. In this paper, we study how to efficiently construct the SFG from a set of VNF requests and map the SFG onto the substrate network while considering the priority dependence between the VNFs. We define the problem of service function graph design and mapping (SFG_PD) and propose an SFG_PD mapping with dependent directional acyclic graph (SFG_DAG) algorithm. The proposed SFG_DAG algorithm can jointly construct the VNFs graph and map VNFs onto the substrate network while minimizing the bandwidth consumption in the substrate network. Our simulation and analysis show that accommodating the VNFs based on the requested bandwidth yields the best performance in terms of the total bandwidth consumption.

Social networks such as Twitter and Facebook have become important sources for users to acquire information. In those social networks, users obtain information from the posts/reposts of their social connections. To acquire information efficiently, users are motivated to connect to users who offer attractive and timely information. In this paper, we study how to effectively optimize social connections to optimize the efficiency of information Acquisition. We define this as the problem of Social Connection Optimization for efficient Information Acquisition (SCOIA). We present our analysis on the information accuracy and timeliness to measure the efficiency of information acquisition. Based on the analysis, a novel User Set Selection (USS) algorithm is then proposed to efficiently solve the SCOIA problem. Our simulations based on the crawled Twitter dataset show that the proposed algorithm can efficiently identify user connections, leading to high information acquisition accuracy, low spam rate and low information acquisition latency.

In this paper, we propose a Reliable Semi-Supervised Learning framework, called ReSSL, for both static and streaming data. Instead of relaxing different assumptions, we do model the reliability of cluster assumption, quantify the distinct importance of clusters (or evolving micro-clusters on data streams), and integrate the cluster-level information and labeled data for prediction with a lazy learning framework. Extensive experiments demonstrate that our method has good performance compared to state-of-the-art algorithms on data sets in both static and real streaming environments.

Assume that each vertex of an arbitrary n-vertex tree T is assigned a nonnegative integer weight. This paper considers partitioning the vertices of tree T into p disjoint clusters such that the total weight of each cluster is at least l and at most u, where l and u are given integers with l