Recent advances in Time Series Foundation Models (TSFMs) have fundamentally revolutionized general time series analysis across domains like finance, retail, weather, and power. However, how to unlock the hidden capacity of general-purpose TSFMs for wearable activity recognition still remains largely unexplored, given severe sensor annotation scarcity and highly heterogeneous sensor data. To address these challenges, we propose DeepSenseMoE—a novel multi-scale convolution-based Mixture of Experts (MoE) module for parameter-efficient fine-tuning of general-purpose TSFMs to sensor-based activity recognition. DeepSenseMoE integrates three key innovations: (1) Multi-scale convolutional experts with different filter sizes responsible for capturing varying sensor contexts; (2) Shared-expert isolation mechanism compressing common activity knowledge into a single shared expert while reducing redundancy among routed experts; and (3) Hierarchical supervised contrastive alignment guiding experts to further learn discriminative activity features. Extensive experiments on three challenging HAR benchmarks demonstrate DeepSenseMoE's superiority, achieving up to 9.5% accuracy gains over state-of-the-art under few-shot and full-supervised settings, with only

In multi-agent systems, explicit cognition of teammates' decision logic serves as a critical factor in facilitating coordination. Communication (i.e., "Tell") can assist in the cognitive development process by information dissemination, yet it is inevitably subject to real-world constraints such as noise, latency, and attacks. Therefore, building the understanding of teammates' decisions without communication remains challenging. To address this, we propose a novel non-communication MARL framework that realizes the construction of cognition through local observation-based modeling (i.e., "Think"). Our framework enables agents to model teammates' active inference process. At first, the proposed method produces three teammate portraits: perception-belief-action. Specifically, we model the teammate's decision process as follows: 1) Perception: observing environments; 2) Belief: forming beliefs; 3) Action: making decisions. Then, we selectively integrate the belief portrait into the decision process based on the accuracy and relevance of the perception portrait. This enables the selection of cooperative teammates and facilitates effective collaboration. Extensive experiments on the SMAC, SMACv2, MPE, and GRF benchmarks demonstrate the superior performance of our method.

Spatial-Temporal Graph (STG) forecasting on large-scale networks has garnered significant attention. However, existing models predominantly focus on short-horizon predictions and suffer from notorious computational costs and memory consumption when scaling to long-horizon predictions and large graphs. Targeting the above challenges, we present FaST, an effective and efficient framework based on heterogeneity-aware Mixture-of-Experts (MoEs) for long-horizon and large-scale STG forecasting, which unlocks one-week-ahead (672 steps at a 15-minute granularity) prediction with thousands of nodes. FaST is underpinned by two key innovations. First, an adaptive graph agent attention mechanism is proposed to alleviate the computational burden inherent in conventional graph convolution and self-attention modules when applied to large-scale graphs. Second, we propose a new parallel MoE module that replaces traditional feed-forward networks with Gated Linear Units (GLUs), enabling an efficient and scalable parallel structure. Extensive experiments on real-world datasets demonstrate that FaST not only delivers superior long-horizon predictive accuracy but also achieves remarkable computational efficiency compared to state-of-the-art baselines. Our source code is available at: https://github.com/yijizhao/FaST.

The proliferation of low-cost, portable sensors has made wearable human activity recognition (HAR) a cornerstone for real-time health monitoring and behavior analysis. However, deploying accurate yet lightweight deep learning models on resource-constrained wearable devices poses a significant challenge for on-device activity recognition. While channel pruning is a common solution to accelerate deep Convolutional Neural Networks (CNNs), existing works often require specialized implementations or pre-trained models, which potentially degrade performance by simply removing an entire channel, limiting their ability to handle complex multimodal sensor inputs. Moreover, lightweight CNN design, particularly the heavy use of 1×1 convolution layers for channel squeezing, remain inefficient for sensor-based HAR, which consume resources without expanding the receptive field due to their pointwise nature. To address these issues, we propose a novel dynamic channel sampling module, Select-and-Fuse (SaF), specifically designed for sensor-based HAR. SaF divides channels into subsets and performs a dynamic, input-dependent selection from them, with the picking decision being made per-time-step based on the input sensor signal activations, allowing for fine-grained feature adaptation to multi-modal sensor signals. While integrated into compact backbones, SaF significantly reduces model size and inference latency while maintaining high accuracy. Extensive evaluations on public UCI-HAR, OPPORTUNITY, WISDM, and UniMiB-SHAR benchmarks confirm a favorable performance-cost trade-off. Crucially, we measure actual inference latency on a Raspberry Pi, proving its practicality for resource-constrained HAR applications. Code will be released.

Lightweight CNNs and Transformers have shown great promise in sensor-based human activity recognition (HAR), yet their structural synergies remain underexplored. This paper bridges this gap by integrating the MetaFormer paradigm—a general architecture abstracted from Transformers that structurally separates token mixing (i.e., self-attention) and channel mixing (i.e., feed-forward networks)—into efficient CNN design. While MetaFormer offers a powerful inductive bias, its standard self-attention mechanism is often computationally intensive for resource-constrained HAR. To address this, we revolutionize the classic MobileNetV3 architecture from a MetaFormer perspective, introducing Rep-MMB, a new family of pure lightweight CNNs. By leveraging structural reparameterization, Rep-MMB decouples multi-branch training-time complexity from efficient single-branch inference, enabling high accuracy with low latency. Evaluations on four public HAR benchmarks show that Rep-MMB outperforms state-of-the-art lightweight models in accuracy and efficiency, with practical validation on embedded devices. We hope that Rep-MMB may serve as a strong baseline to inspire future edge-deployed HAR research.

Human Activity Recognition (HAR) plays a pivotal role in ubiquitous computing. However, it remains constrained by the challenge of balancing fine-grained temporal modeling with real-time efficiency on resource-limited devices. While Transformer-based models excel at capturing long-range dependencies, they suffer from high computational costs, limiting their applicability on resource-constrained devices. Linear attention mechanisms improve efficiency but often discard negative signals and produce overly smooth, high-entropy attention distributions, impairing the extraction of fine-grained patterns and degrading classification accuracy in complex scenarios. In this work, we present ActiFormer, a novel sign-aware linear attention framework tailored for sensor-based HAR to overcome these limitations. To preserve bidirectional signal dynamics, we introduce Sign-Aware Attention, which explicitly models both same-sign and cross-sign interactions between queries and keys, effectively retaining negative signals crucial for accurate recognition. Furthermore, we propose a learnable entropy-scaling function that compensates for the exponential scaling effect lost in linear attention, originally provided by softmax, solving the high-entropy attention weight issue by amplifying the importance of critical temporal points. Extensive experiments on four benchmark HAR datasets demonstrate that ActiFormer consistently outperforms CNNs, standard Transformers, and state-of-the-art linear attention models, both in accuracy and efficiency. Its lightweight design supports real-time inference on edge devices such as the Raspberry Pi 5, highlighting its practical deployability in real-world applications.

Transformer models have demonstrated superior capability in capturing long-range temporal dependencies crucial for Sensor-Based Human Activity Recognition (HAR). However, the quadratic computational complexity inherent to the Softmax-Attention mechanism significantly impedes their deployment on resource-constrained wearable devices and real-time streaming tasks. To address this, we propose a novel Linear Transformer with Taylor Series Attention specifically tailored for the HAR domain, named TSA-Former. It leverages the first-order Taylor expansion to approximate the Softmax-Attention and utilizes the norm-preserving mapping to approximate the high-order non-linear information, resulting in a linear computational complexity. In addition, TSA-Former integrates a multi-branch architecture featuring multi-scale patch embedding, which enables the model to dynamically capture multi-scale temporal features while minimizing overhead. Experimental results across four public HAR benchmarks, namely UniMiB-SHAR, UCI-HAR, WISDM, and OPPORTUNITY, demonstrate that TSA-Former achieves state-of-the-art (SOTA) accuracy and efficiency, outperforming conventional Transformers and existing linear-attention models. Deployment experiments conducted on the Raspberry Pi 5 platform further validate the model’s superior low-latency and minimal power consumption profile, confirming its robust suitability for real-world embedded HAR applications. Code will be released.

Human Activity Recognition (HAR) aims to classify human behaviors from large-scale sensor data. A key challenge is to achieve high recognition accuracy while maintaining low computational cost. Recent advances such as Mamba address this by enabling long-range dependency modeling with subquadratic computational complexity, thus achieving strong representational capacity at reduced cost. However, when directly applied to HAR tasks, lightweight Mamba-based backbones often underperform compared to conventional CNN and Transformer architectures. To investigate this gap, we perform detailed temporal and spectral analyses, revealing that Mamba exhibits an inherent bias towards low-frequency components. In contrast, HAR sensor signals typically comprise a mixture of both high- and low-frequency information, both of which are crucial for accurate activity recognition. To address this limitation, we propose Machar, a novel lightweight MAmba-Convolution Hybrid ARchitecture specifically designed for HAR. Instead of relying solely on global modeling, Machar introduces a dedicated FreqDecoupler that decomposes sensor signals into high- and low-frequency components, enabling each to be processed by the most appropriate mechanism. Furthermore, we propose a frequency scheduling strategy that dynamically adjusts channel capacity allocation across network stages, effectively combining the local feature extraction capability of CNNs with Mamba’s global modeling strength. Extensive experiments on three widely used HAR benchmarks, namely USC-HAD, UCI-HAR, and UniMiB-SHAR, show that Machar consistently outperforms existing methods, achieving impressive accuracy while preserving a favorable computational footprint, which underscore the effectiveness and scalability of Machar for real-world HAR applications.

Lightweight convolution-based neural networks (CNNs) are well suited for sensor-based human activity recognition (HAR) applications on resource-constrained edge devices with faster inference speed. However, the convolutional kernels are often limited to a small window range, which can only capture local details in time series sensor data, thus preventing further performance boost. Though Introducing self-attention into convolution can help to handle long-range dependence well, it might significantly slow down actual activity inference speed, due to high computational cost. In this paper, we introduce a new learning paradigm (star operation) and then present a lightweight Dual-Branch High-Order Interactions (DbHoi) block, which is computationally friendly for mobile HAR deployment. The proposed DbHoi block may implicitly transform raw sensor inputs into high-dimensional non-linear features, but actually operate in a low-dimensional feature space (analogs to the design principle of polynomial kernel tricks), without incurring extra computational overhead. Extensive experiments are conducted on three public HAR benchmarks including UCI-HAR, UniMiB-SHAR, and OPPORTUNITY, which demonstrate that our suggested DbHoi can consistently surpass various meticulously designed lightweight networks such as MobileNet, ShuffleNet, and GhostNet. Detailed ablation studies, visualizing representations, and on-device latency analyses further validate our insights with regards to the star operation, while underscoring its practical merit in real-world HAR deployment.

Inspired by recent success of foundation models in vision and language domains, time series foundation models (TSFMs) have garnered increasing attention in general time series analysis tasks like finance, weather, healthcare, and power. However, given high heterogeneity and severe annotation scarcity in time series sensor data, how to unlock the potential of large-scale general-purpose TSFMs for downstream activity recognition tasks remains yet unexplored? This paper makes the first attempt to address this timely challenge by adapting the self-supervised pre-trained TSFM (i.e., MOMENT) to few-shot activity recognition. We introduce a simple and efficient Sensor-Prompt Tuning (SPT) strategy, which employs multiple convolution-based sensor-friendly filters with a gating mechanism to act as learnable soft prompts, which can dynamically adapt sensor input space to the frozen TSFM backbone, effectively bridging domain gap between pre-training general time series data with wearable sensor stream. Extensive experiments across three public activity recognition benchmarks demonstrate that our SPT achieves up to 15.5% performance gains over existing state-of-the-art baselines under few-shot scenarios, while considerably outperforming other mainstream fine-tuning strategies with smaller than 1% of backbone parameters. Practical cloud-edge inference latencies are measured. This work offers a new prompt-tuning perspective on how to adapt pre-trained TSFMs for wearable activity recognition tasks. Code will be released.

Communication has been widely employed to enhance multi-agent collaboration. Previous research has typically assumed delay-free communication, a strong assumption that is challenging to meet in practice. However, real-world agents suffer from channel delays, receiving messages sent at different time points, termed Asynchronous Communication, leading to cognitive biases and breakdowns in collaboration. This paper first defines two communication delay settings in MARL and emphasizes their harm to collaboration. To handle the above delays, this paper proposes a novel framework, Communication Delay-Tolerant Multi-Agent Collaboration (CoDe). At first, CoDe learns an intent representation as messages through future action inference, reflecting the stable future behavioral trends of the agents. Then, CoDe devises a dual alignment mechanism of intent and timeliness to strengthen the fusion process of asynchronous messages. In this way, agents can extract the long-term intent of others, even from delayed messages, and selectively utilize the most recent messages that are relevant to their intent. Experimental results demonstrate that CoDe outperforms baseline algorithms in three MARL benchmarks without delay and exhibits robustness under fixed and time-varying delays.

Old photo restoration is a challenging task due to diverse and compound degradations, complex structural damage, and the lack of real paired training data. To address these issues, we present a novel diffusion-based framework that integrates contrastive learning and frequency-aware modeling to effectively restore old photographs. Specifically, we design a category-guided contrastive encoder to align the latent distributions of synthetic and real old photos, effectively bridging the domain gap. Based on this aligned latent space, a diffusion model is employed to iteratively generate high-quality restorations with strong semantic and structural consistency. Additionally, we introduce a Discrete Cosine Histogram Attention (DCH) block to jointly capture structured (e.g., holes, scratches, blotches) and unstructured (e.g., noise, blur, fading, low resolution) degradations in the frequency domain. Without relying on real paired data, our method generalizes well to real-world old photos. Extensive experiments demonstrate that our approach consistently outperforms state-of-the-art methods in both synthetic and real-world scenarios. The code is available at https://github.com/Marvel-LG/Code.

Continual Multi-Agent Reinforcement Learning (Co-MARL) requires agents to address catastrophic forgetting issues while learning new coordination policies with the dynamics team. In this paper, we delve into the core of Co-MARL, namely Relation Patterns, which refer to agents’ general understanding of interactions. In addition to generality, relation patterns exhibit task-specificity when mapped to different action spaces. To this end, we propose a novel method called General Relation Patterns-Guided Task-specific Decision-Maker (RPG). In RPG, agents extract relation patterns from dynamic observation spaces using a relation capturer. These task-agnostic relation patterns are then mapped to different action spaces via a task-specific decision-maker generated by a conditional hypernetwork. To combat forgetting, we further introduce regularization items on both the relation capturer and the conditional hypernetwork. Results on SMAC and LBF demonstrate that RPG effectively prevents catastrophic forgetting when learning new tasks and achieves zero-shot generalization to unseen tasks.

Diffusion transformer models have made significant advancements in the generation of visual media. Despite their impressive capabilities, these models often face considerable computational overhead, mainly stemming from their sequential denoising process and large size. Traditional methods for compressing diffusion models generally require extensive retraining, which introduces both cost and feasibility challenges. Feature caching methods have been proposed to accelerate diffusion transformers by caching features from previous timesteps and reusing them in subsequent timesteps, without necessitating additional training. Nevertheless, previous caching techniques fail to account for the varying sensitivities of different tokens to feature caching. Consequently, indiscriminate caching of certain tokens may lead to significant degradation in overall generation quality. In this work, we introduce a novel caching method that enables diffusion models to leverage image structural information, which is an essential component of the diffusion process, while adaptively selecting the most appropriate tokens for caching. Experiments on diverse downstream tasks reveal that the proposed method enhances generation efficiency while maintaining high visual quality. Moreover, it provides the advantages of being plug-and-play, with no requirement for additional training.

Recently, deep convolutional neural networks (CNNs) have achieved outstanding success in sensor-based human activity recognition (HAR) scenario, but at the cost of huge computational complexity, thereby restricting their practical deployment on resource-limited wearable devices. This may be partly attributed to static nature of most existing CNNs, which process all activity samples uniformly, resulting in structural and data redundancy. Comparing to static networks, one promising strategy is to accelerate activity inference by exploiting structural redundancy within deep CNNs, which selectively activates computation units such as convolution channels while handling different samples. The other promising strategy is to explore spatiotemporal redundancy by concentrating computational effort on the most informative regions of sensor data. How to simultaneously leverage structural and data redundancy still remains largely overlooked. In this article, from a new perspective of exploring both structural and spatiotemporal redundancy, we introduce an efficient spatiotemporal-structural masker network (SSMNet) for activity recognition. It utilizes a dual-mask mechanism to make dynamic, sample-specific decisions, thereby accelerating activity inference. The spatiotemporal-structural masker integrates spatiotemporal and structural decisions through masks, dynamically allocating computational resources based on input with minimal overhead. Extensive experiments on three public HAR benchmark datasets, namely, WISDM, UniMiB-SHAR, and PAMAP2. SSMNet is guided by a high-accuracy static model, allowing it to reduce computational costs while maintaining state-of-the-art performance. For example, comparing to static baselines, it may reduce nearly 40% FLOPs with an accuracy drop smaller than 1%, across all three datasets The detailed analyses affirm that our method can strike an optimal tradeoff between accuracy and efficiency.

Mobile crowdsensing service (MCS) providers recruit users to complete data collection tasks by rewarding the users to obtain greater revenue. Therefore, maximizing revenue is a focus of the MCS provider. This article expresses this problem as a revenue maximization programming model with budget constraints and designs a reverse-auction mechanism based on the attention model to solve the task allocation and pricing problems. Specifically, we convert the programming model under multiple constraints into an augmented Lagrangian function, optimally solve it through a multilayer neural network on the basis of the attention interactive framework, and finally output the allocation and payment solution. Our design guarantees that the mechanism meets economic criteria such as truthfulness, individual rationality, and budget feasibility. Combining the revenue-optimal reverse-auction mechanism with deep learning provides a new approach to mechanism design. Compared with existing methods, our solution achieves very good results in terms of service provider revenue and generalization experiments.

Various Web API Recommendation (AR) techniques have assisted developers in efficiently identifying suitable APIs for mashup creation. With the emergence of large language models (LLMs), there has been increasing interest in leveraging LLMs for recommender systems. Although several approaches have attempted to utilize LLMs by framing recommendations as prompts, this approach is not ideally suited for AR due to fundamental differences in the training processes of LLMs and AR models. Consequently, it's crucial to conduct further research to identify effective applications of LLMs in AR. To this end, we propose a novelLLM-based generative solution forAPIRecommendation (LLMAR) that combines instruction learning of multitask and multistage Low-Rank Adaptation fine-tuning based on LLaMA models. Experimental results on the ProgrammableWeb dataset show that LLMAR significantly outperforms representative methods in regular and data-limited scenarios.

During recent years, dynamic early-exit has provided a promising paradigm to improve the computational efficiency of deep neural networks by constructing multiple classifiers to let easy samples exit at shallow layers while avoiding redundant computations at deep exits, which has been seldom explored in the context of latency-aware human activity recognition (HAR) deployed on wearable devices. Particularly, most existing early-exit strategies have always treated all activity samples equally at each exit during training, which ignore such dynamic early-exit behavior at test-time, causing a potential mismatch between training and test. Intuitively, easy activity samples that often exit earlier at test-time should place more emphasis on the training loss of shallow classifiers, while hard activity samples should contribute more to the training loss of deep classifiers. To bridge this gap, this paper introduces a sample-reweighting approach for efficient activity inference, which employs a weight-predicting network to reweight the training loss of different activity samples at every exit. From a perspective of meta learning, a new optimization objective function is designed to jointly optimize both weight-predicting network and backbone network. We perform extensive experiments on three popular HAR benchmarks including UCI-HAR, WISDM, and UniMiB-SHAR, which demonstrate that while incorporating such test-time early-exit behavior into conventional training pipeline, it can consistently improve the accuracy-efficiency trade-offs under budgeted batch classification and anytime prediction patterns. Moreover, our approach has a natural advantage in handing class-imbalance HAR problem. Detailed ablation studies, visualized illustrations, and real hardware deployment are provided to support our statement.

Mobile crowdsensing service (MCS) providers recruit users to complete data collection tasks with an incentive mechanism. How to maximize the utility of service providers has long been a popular topic in MCS research. Applying the existing reverse auction mechanism to an MCS may result in excessively high payments, thereby reducing the utility of the MCS provider. The affine maximizer auction (AMA) mechanism increases the revenue of service providers and meets dominant-strategy incentive-compatible (DSIC) characteristics. However, the AMA mechanism is a forward auction mechanism and cannot be applied to MCSs. Inspired by the AMA mechanism, this paper innovatively proposes a reverse affine maximizer auction (RAMA) mechanism to solve the task allocation problem of MCSs, effectively improving the MCS provider utility. Specifically, we construct a RAMA theoretical model and prove that the mechanism satisfies DSIC characteristics. For the discrete MCS task allocation problem, we use the reverse virtual valuation combinatorial auction (RVVCA) mechanism, a subclass of RAMA, to design a random mechanism RVVCA$^{t}$and prove that the RVVCA$^{t}$has a logarithmic approximate ratio. For the differentiable MCS task allocation problem, we use the deep learning transformer framework to design RAMANet, which can fit an exponential number of allocation solutions and output the optimal allocation and payment. We experimentally compare the algorithms of the RAMA family we propose, which use affine maximization, with existing state-of-the-art algorithms, demonstrating that the proposed algorithms significantly improve MCS provider utility.

Abstract Due to the richness of natural image semantics, natural image colourisation is a challenging problem. Existing methods often suffer from semantic confusion due to insufficient semantic understanding, resulting in unreasonable colour assignments, especially at the edges of objects. This phenomenon is referred to as colour bleeding. The authors have found that using the self‐attention mechanism benefits the model's understanding and recognition of object semantics. However, this leads to another problem in colourisation, namely dull colour. With this in mind, a Position‐Spatial Attention Network(PSANet) is proposed to address the colour bleeding and the dull colour. Firstly, a novel new attention module called position‐spatial attention module (PSAM) is introduced. Through the proposed PSAM module, the model enhances the semantic understanding of images while solving the dull colour problem caused by self‐attention. Then, in order to further prevent colour bleeding on object boundaries, a gradient‐aware loss is proposed. Lastly, the colour bleeding phenomenon is further improved by the combined effect of gradient‐aware loss and edge‐aware loss. Experimental results show that this method can reduce colour bleeding largely while maintaining good perceptual quality.

During recent years, deep neural networks have achieved outstanding success in sensor-based human activity recognition (HAR). Particularly, dynamic convolution has emerged as a promising solution to accelerate activity inference of deep networks on mobile devices. Exploiting spatial redundancy, such a dynamic strategy can adaptively sample the salient areas of interest over sensor feature maps while skipping unimportant locations to avoid computational expenditure on activity-irrelevant disturbing areas. Despite theoretic efficiency, it has to rely on a binary-valued mask combined with element-wise multiplication, which potentially incurs noncontiguous memory access while performed at the finest granularity. To the best of our knowledge, most existing HAR literatures have always adopted hardware-agnostic FLOPs as an indicator to guide the algorithm design, lacking delay-aware considerations about scheduling strategy and specific hardware characteristic. In this article, we propose a delay-aware coarse-grained dynamic convolutional network called DACDNet to bridge the gap between theoretical FLOPs and realistic delay, which is highly challenging but less explored in ubiquitous HAR environments. Instead of theoretic FLOPs, we introduce a novel delay prediction model to guide the HAR algorithm design while simultaneously considering the scheduling strategy on various hardware platforms, especially multicore processors like the edge GPU devices. Experiments on multiple HAR benchmarks, including WISDM, UniMiB-SHAR, and PAMAP2 demonstrate that our approach can significantly accelerate activity inference without sacrificing accuracy.

During recent years, deep convolutional neural networks have demonstrated dominant performance in human activity recognition (HAR) using wearable sensors. However, they often come at high computational cost when fueled with fixed-length sliding window. This article primarily aims to accelerate activity inference from a novel perspective of reducing temporal redundancy in sensor data. Inspired by the fact that not all time intervals within a window are activity-relevant, we formulate the activity prediction problem as a dynamic inference process by continuously attending to a sequence of small activity-discriminative intervals, which are selected from an original window by progressively predicting the discriminative importance of each interval with an interpretable interval proposal network. The dynamic process can adaptively decide when to halt for each individual sample, which considerably avoids excessive computation by letting “easy” activity exit as early as possible while progressively focusing on small salient intervals for “hard” activity. Given a limited budget, the accuracy-cost tradeoff can be flexibly and precisely controlled via tuning confidence thresholds online without requiring to be retrained from scratch—a practical requirement in real-world HAR applications. Extensive experiments on several standard benchmarks including University of California-Irvine-Human Activity Recognition (UCI-HAR), wireless sensor data mining (WISDM), University of Southern California-Human Activity Dataset (USC-HAD), and Weakly Labeled dataset demonstrate that our dynamic inference process significantly outperforms previous static methods according to theoretical and practical computational efficiency.

Self-supervised Human Activity Recognition (HAR) has been gradually gaining a lot of attention in ubiquitous computing community. Its current focus primarily lies in how to overcome the challenge of manually labeling complicated and intricate sensor data from wearable devices, which is often hard to interpret. However, current self-supervised algorithms encounter three main challenges: performance variability caused by data augmentations in contrastive learning paradigm, limitations imposed by traditional self-supervised models, and the computational load deployed on wearable devices by current mainstream transformer encoders. To comprehensively tackle these challenges, this paper proposes a powerful self-supervised approach for HAR from a novel perspective of denoising autoencoder, the first of its kind to explore how to reconstruct masked sensor data built on a commonly employed, well-designed, and computationally efficient fully convolutional network. Extensive experiments demonstrate that our proposed Masked Convolutional AutoEncoder (MaskCAE) outperforms current state-of-the-art algorithms in self-supervised, fully supervised, and semi-supervised situations without relying on any data augmentations, which fills the gap of masked sensor data modeling in HAR area. Visualization analyses show that our MaskCAE could effectively capture temporal semantics in time series sensor data, indicating its great potential in modeling abstracted sensor data. An actual implementation is evaluated on an embedded platform.

Despite strong representation ability, deep convolutional neural networks (CNNs) are largely hindered in practical human activity recognition (HAR) deployment due to high computational cost, which is often unaffordable on resource-limited wearable devices. In this article, to bridge the gap between on-device HAR and deep learning, we present a collaborative compression scheme to reduce the runtime of HAR with an acceptable performance degradation, which combines channel pruning and tensor decomposition to simultaneously handle sparsity and low-rankness when fully considering mutual interference in one network consisting of efficient 1-dimensional convolutional kernels. Our method includes two main stages. Concretely, given a target compression ratio, a global compression ratio decision optimization is first performed to automatically decide per-layer compression ratio by measuring compression sensitivity, without requiring labor-exhaustive human intervention. Then a multi-step collaborative compression is iteratively implemented to remove the least important compression unit based on an improved importance metric until the per-layer target compression ratio is attained. Extensive experiments on multiple HAR benchmarks show that our approach considerably outperforms previous compression strategies. For example, it can achieve around 50% FLOPs reduction with only an accuracy drop of 0.25% and 0.15% on UCI-HAR and PAMAP2, respectively. Actual implementation is evaluated on an embedded platform.

Mobile crowdsensing services are divided into two categories: opportunistic and participatory. In opportunistic mobile crowdsensing services, users do not need to specify the crowdsensing tasks to be completed. Compared with participatory crowdsensing services, the application scope is wider and more user-friendly. In participatory crowdsensing, the service provider assumes that the user can successfully complete the data collection task. However, such an approach cannot work in an opportunistic crowdsensing service because in opportunistic crowdsensing, the user’s execution of the task is uncertain, which brings great challenges to the quality of the crowdsensing service. This article is based on the assumption of the user coverage probability model and transforms the opportunistic mobile crowdsensing value maximization problem into an ordered submodularity value function model with budget constraints. This model is also good at representing participatory crowdsourcing problems. To the best of our knowledge, this is the first study to apply the ordered submodularity feature to a mobile crowdsensing service. Furthermore, we combine the properties of ordered submodular and auction models and propose an ordered submodularity-proportional share mechanism (O-PSM) to solve the allocation and payment problems in opportunistic mobile crowdsensing services. Specifically, in the allocation stage, the winning users are selected based on the proportional share threshold, and in the payment stage, the payment price for the winning users is designed based on critical value theory. We prove that the mechanism satisfies the economic characteristics of individual rationality, truthfulness, and budget feasibility. In the experimental section, the mechanism design based on ordered submodularity is shown to enable the service provider to obtain a higher value and a lower payment.

Graph neural network (GNN) is a robust model for processing non-Euclidean data, such as graphs, by extracting structural information and learning high-level representations. GNN has achieved state-of-the-art recommendation performance on collaborative filtering (CF) for accuracy. Nevertheless, the diversity of the recommendations has not received good attention. Existing work using GNN for recommendation suffers from the accuracy-diversity dilemma, where slightly increases diversity while accuracy drops significantly. Furthermore, GNN-based recommendation models lack the flexibility to adapt to different scenarios' demands concerning the accuracy-diversity ratio of their recommendation lists. In this work, we endeavor to address the above problems from the perspective of aggregate diversity, which modifies the propagation rule and develops a new sampling strategy. We propose graph spreading network (GSN), a novel model that leverages only neighborhood aggregation for CF. Specifically, GSN learns user and item embeddings by propagating them over the graph structure, utilizing both diversity-oriented and accuracy-oriented aggregations. The final representations are obtained by taking the weighted sum of the embeddings learned at all layers. We also present a new sampling strategy that selects potentially accurate and diverse items as negative samples to assist model training. GSN effectively addresses the accuracy-diversity dilemma and achieves improved diversity while maintaining accuracy with the help of a selective sampler. Moreover, a hyper-parameter in GSN allows for adjustment of the accuracy-diversity ratio of recommendation lists to satisfy the diverse demands. Compared to the state-of-the-art model, GSN improved R @20 by 1.62%, N @20 by 0.67%, G @20 by 3.59%, and E @20 by 4.15% on average over three real-world datasets, verifying the effectiveness of our proposed model in diversifying overall collaborative recommendations.

Dynamic quality of service (QoS) measurement is crucial for discovering services and developing online service systems. Collaborative filtering-based approaches perform dynamic QoS prediction by incorporating temporal information only but never consider the dynamic network environment and suffer from poor performance. Considering different service invocation routes directly reflect the dynamic environment and further lead to QoS fluctuations, we coin the problem of Dynamic QoS Prediction (DQP) with Intelligent Route Estimation (IRE) and propose a novel framework named IRE4DQP. Under the IRE4DQP framework, the dynamic environment is captured by Network Status Representation, and the IRE is modeled as a Markov decision process and implemented by a deep learning agent. After that, the DQP is achieved by a specific neural model with the estimated route as input. Through collaborative training with reinforcement and inverse reinforcement learning, eventually, based on the updated representations of the network status, IRE learns an optimal route policy that matches well with observed QoS values, and DQP achieves accurate predictions. Experimental results demonstrate that IRE4DQP outperforms SOTA methods on the accuracy of response-time prediction by 5.79–31.34% in MAE, by 1.29–20.18% in RMSE, and by 4.43–27.73% in NMAE and with a success rate of nearly 45% on finding routes.

Accurate and reliable service quality prediction has become a key issue in service recommendation and network measurement scenarios. However, traditional methods for time-aware QoS prediction face two main challenges: (I) data sparsity makes it difficult to estimate and recover global information from the limited known data; (II) shallow learning models struggle to represent the intricate relationships between objects, and thus suffer poor prediction performance. To this end, we propose a time-aware QoS prediction framework that combines the merits of graph modeling, graph representation learning, and contrastive learning. First, a novel graph schema is proposed to capture the complex interactions between user-service-slots. Then, a prediction model is developed leveraging a graph convolutional network to learn the node representations by aggregating feature information from neighboring nodes. Finally, a novel contrastive learning strategy is used to improve the robustness of node representation. Experimental results on a large-scale dataset demonstrated that our proposed method significantly outperforms the state-of-the-art prediction methods on response time and throughput prediction tasks.

Machine Learning (ML) workflows are increasingly deployed on serverless computing platforms to benefit from their elasticity and fine-grain pricing. Proper resource allocation is crucial to achieve fast and cost-efficient execution of serverless ML workflows (specially for hyperparameter tuning and model training). Unfortunately, existing resource allocation methods are static, treat functions equally, and rely on offline prediction, which limit their efficiency. In this paper, we introduce CE-scaling – a Cost-Efficient autoscaling framework for serverless ML work-flows. During the hyperparameter tuning, CE-scaling partitions resources across stages according to their exact usage to minimize resource waste. Moreover, it incorporates an online prediction method to dynamically adjust resources during model training. We implement and evaluate CE-scaling on AWS Lambda using various ML models. Evaluation results show that compared to state-of-the-art static resource allocation methods, CE-scaling can reduce the job completion time and the monetary cost by up to 63% and 41% for hyperparameter tuning, respectively; and by up to 58% and 38% for model training.

During the past decade, human activity recognition ( HAR ) using wearable sensors has become a new research hot spot due to its extensive use in various application domains such as healthcare, fitness, smart homes, and eldercare. Deep neural networks, especially convolutional neural networks ( CNNs ), have gained a lot of attention in HAR scenario. Despite exceptional performance, CNNs with heavy overhead is not the best option for HAR task due to the limitation of computing resource on embedded devices. As far as we know, there are many invalid filters in CNN that contribute very little to output. Simply pruning these invalid filters could effectively accelerate CNNs , but it inevitably hurts performance. In this article, we first propose a novel CNN for HAR that uses filter activation. In comparison with filter pruning that is motivated for efficient consideration, filter activation aims to activate these invalid filters from an accuracy boosting perspective. We perform extensive experiments on several public HAR datasets, namely, UCI-HAR ( UCI ), OPPORTUNITY ( OPPO ), UniMiB-SHAR ( Uni ), PAMAP2 ( PAM2 ), WISDM ( WIS ), and USC-HAD ( USC ), which show the superiority of the proposed method against existing state-of-the-art ( SOTA ) approaches. Ablation studies are conducted to analyze its internal mechanism. Finally, the inference speed and power consumption are evaluated on an embedded Raspberry Pi Model 3 B plus platform.

Developers need to reuse web services and create mashups suitable for various scenarios. Currently, it relies on the developer’s adequate domain knowledge to be able to find services and verify their compatibility. Although service recommendation systems already exist to assist them, inexperienced developers may not be able to adequately express their requirements, resulting in inappropriate and incompatible recommendations. To tackle this problem, we define a service-keyword correlation graph (SKCG) to capture the relationship between services and keywords, and the compatibility among services. Then, we propose keyword-based deep reinforced Steiner tree search (K-DRSTS) to recommend services for mashup creation. K-DRSTS models the task of service discovery as a Steiner tree search problem against SKCG. Leveraging deep reinforcement learning, K-DRSTS provides an efficient solution for solving the NP-hard search problem of the Steiner tree. Extensive experiments on real-world data sets have shown the effectiveness of K-DRSTS.

In order to overcome the low robustness and weak generalization in existing deep autoencoder (AE) for soft sensor modeling, a novel feature-disentangled AE (FDAE) integrating residual network (Resnet) (FDAE-Resnet) is proposed in this article. Different from the traditional deep AE that only can learn entangled features, the FDAE can obtain disentangled multisource features including trend features, periodic features and spatial features by a new trend-periodic long short-term memory (TPLSTM) and a novel dynamic self-attention convolutional neural network (DSACNN). The trend and periodic signals decomposed from input variables are fed into the TPLSTM to learn trend and periodic features in time and frequency domain, respectively. Then, the DSACNN is utilized to capture dynamic spatial features in spatial domain by adding a new attention mechanism. Moreover, disentangled multisource features are obtained by concatenating trend features, periodic features and spatial features together. Finally, the Resnet is utilized to build the soft sensor model by establishing the relationship between disentangled multisources features and outputs. To illustrate the effectiveness and superiority of the proposed method, the FDAE-Resnet is applied in the actual polypropylene process industry for melt index modeling. The experiment results show that compared with other state-of-the-art methods, the FDAE-Resnet can reduce the root mean square error by 26.2% and the mean absolute percentage error by 38.2% on average in the changed working conditions, respectively.

Federated Learning (FL) has recently attracted great interest in sensor-based human activity recognition (HAR) tasks. However, in real-world environment, sensor data on devices is non-independently and identically distributed (Non-IID), e.g., activity data recorded by most devices is sparse, and sensor data distribution for each client may be inconsistent. As a result, the traditional FL methods in the heterogeneous environment may incur a drifted global model that causes slow convergence and a heavy communication burden. Although some FL methods are gradually being applied to HAR, they are designed for overly ideal scenarios and do not address such Non-IID problem in the real-world setting. It is still a question whether they can be applied to cross-device FL. To tackle this challenge, we propose ProtoHAR, a prototype-guided FL framework for HAR, which aims to decouple the representation and classifier in the heterogeneous FL setting efficiently. It leverages the global prototype to correct the activity feature representation to make the prototype knowledge flow among clients without leaking privacy while solving a better classifier to avoid excessive drift of the local model in personalized training. Extensive experiments are conducted on four publicly available datasets: USC-HAD, UNIMIB-SHAR, PAMAP2, and HARBOX, which are collected in both controlled environments and real-world scenarios. The results show that compared with the state-of-the-art FL algorithms, ProtoHAR achieves the best performance and faster convergence speed in HAR datasets.

Recent years have witnessed great success of deep convolutional networks in sensor-based human activity recognition (HAR), yet their practical deployment remains a challenge due to the varying computational budgets required to obtain a reliable prediction. This article focuses on adaptive inference from a novel perspective of signal frequency, which is motivated by an intuition that low-frequency features are enough for recognizing "easy" activity samples, while only "hard" activity samples need temporally detailed information. We propose an adaptive resolution network by combining a simple subsampling strategy with conditional early-exit. Specifically, it is comprised of multiple subnetworks with different resolutions, where "easy" activity samples are first classified by lightweight subnetwork using the lowest sampling rate, while the subsequent subnetworks in higher resolution would be sequentially applied once the former one fails to reach a confidence threshold. Such dynamical decision process could adaptively select a proper sampling rate for each activity sample conditioned on an input if the budget varies, which will be terminated until enough confidence is obtained, hence avoiding excessive computations. Comprehensive experiments on four diverse HAR benchmark datasets demonstrate the effectiveness of our method in terms of accuracy-cost tradeoff. We benchmark the average latency on a real hardware.

During recent years, channel attention has attracted great interest in deep learning community. Despite significant success, it has been rarely exploited in ubiquitous human activity recognition (HAR) scenario. To decrease computational overhead, the channel attention often uses global averaging pooling (GAP) to compress each channel into a simple scalar. It is well known that GAP is equal to the lowest frequency component. Despite obvious lightweight advantage, such compression process inevitably causes severe information loss. In this paper, we propose a novel multi-frequency channel attention framework for activity recognition tasks. Considering various sensing frequencies of human activities, an intuition solution is to convert the time series from time domain to frequency domain. Instead of GAP, the discrete cosine transform (DCT) is used to compress channels. We prove that GAP can be seen as a special case of DCT, which uses the lowest frequency component only and leaves out all other frequency components unused. DCT is able to better compress channels by fully exploiting other frequency components discarded by GAP. Despite multiple frequency components used, each channel will still be represented by a scalar in order to maintain the same computational overhead. Using two frequency screening criteria, our method is able to achieve state-of-the-art results on four benchmark HAR datasets. Extensive ablation studies are conducted, which provides a better interpretability of deep model behaviors. Finally, actual inference is evaluated on an embedded platform.

Recently, human activity recognition (HAR) that uses wearable sensors has become a research hotspot because its wide applications in real-world scenarios. Essentially, HAR can be treated as multi-channel time series classification problem, where different channels may come from heterogeneous sensor modalities. Deep learning, especially convolutional neural networks (CNNs) have made breakthroughs in ubiquitous HAR scenario. Various normalization methods enable layers of networks to learn more independently by normalizing hybrid sensor features. However, normalization tends to produce a channel collapse phenomenon, where many channels generates tiny values. Most channels are inhibited and contribute very little to output. As a result, the network has to rely on only a few valid channels, which inevitably impair the generality ability. In this paper, we provide an alternative called Channel Equalization to reactivate these inhibited channels by performing whitening or decorrelation operation, which compels all channels to contribute more or less to feature representation. Extensive experiments are conducted on several public HAR benchmarks, which indicate that the proposed method significantly surpasses recent SOTA at negligible computational overhead. To our knowledge, the Channel Equalization is for the first time to be applied in multimodal HAR scenario. Finally, the actual operation is evaluated on an embedded platform.

State-of-the-art image style transfer methods have achieved impressive results by using neural networks. However, neural style transfer (NST) methods either ignore the local details of the style image by using the global statistics for style modeling or cannot fully use shallow features of neural networks, leading to the synthesized image having fewer details. In this study, we proposed a new patch-based style transfer method that directly operates in the image pixel domain without using any neural networks, achieving fascinating style transfer results with rich image details. The proposed method was derived from classic texture synthesis methods. Most previous methods rely on nearest neighbor search (NNS) for patch matching. However, this greedy strategy cannot guarantee the similarity of patch distributions between the synthesized image and the style image, which limits the expressiveness of textures. We solved this problem by proposing an optimal patch matching algorithm formed on the Optimal Transport (OT) theory, which theoretically guarantees the similarity of the patch distributions and gives a flexible style modeling method. Various qualitative and quantitative experiments demonstrated that the proposed method achieves better synthesized results than state-of-the-art style transfer methods, including NST and classic methods based on texture synthesis.

Multiview subspace clustering has turned into a promising technique due to its encouraging ability to discover the underlying subspace structure. In recent studies, a lot of subspace clustering methods have been developed to strengthen the clustering performance of multiview data, but these methods rarely consider simultaneously the nonlinear structure and multilevel representation (MLR) information in multiview data as well as the data distribution of latent representation. To address these problems, we develop a new Multiview Subspace Clustering with MLRs and Adversarial Regularization (MvSC-MRAR), where multiple deep auto-encoders are utilized to model nonlinear structure information of multiview data, multiple self-expressive layers are introduced into each deep auto-encoder to extract multilevel latent representations of each view data, and diversity regularizations are designed to preserve complementary information contained in different layers and different views. Furthermore, a universal discriminator based on adversarial training is developed to enforce the output of each encoder to obey a given prior distribution, so that the affinity matrix for spectral clustering (SPC) is more realistic. Comprehensive empirical evaluation with nine real-world multiview datasets indicates that our proposed MvSC-MRAR achieves significant improvements than several state-of-the-art methods in terms of clustering accuracy (ACC) and normalized mutual information (NMI).

End-to-end QoS measurement plays an indispensable role in the decision-making of cloud services and IoT services. Many efforts have paid on developing QoS prediction approaches in the past decade leveraging the principle of collaborative filtering. But there remain many challenging issues concerning multi-task prediction requirements, feature selection for heterogeneous prediction tasks, and model training. To this end, we propose an effective personalized service QoS prediction method from the perspective of multi-task learning, named PMT. PMT consists of specially-designed feature selection components and a multi-step model training strategy. The feature selection method leverages the principle of multi-expert decision-making and self-attention mechanism. The multi-step model training enables a weight-free configuration for parallel prediction tasks. Experimental results on a large dataset with two tasks and a small dataset with three tasks demonstrate that PMT is superior to the state-of-the-art QoS prediction methods.

Automated code generation is a longstanding challenge in both communities of software engineering and artificial intelligence. Currently, some works have started to investigate the functional correctness of code generation, where a code snippet is considered correct if it passes a set of test cases. However, most existing works still model code generation as text generation without considering program-specific information, such as functionally equivalent code snippets and test execution feedback. To address the above limitations, this paper proposes a method combining program analysis with deep learning for neural code generation, where functionally equivalent code snippets and test execution feedback will be considered at the training stage. Concretely, we firstly design several code transformation heuristics to produce different variants of the code snippet satisfying the same functionality. In addition, we employ the test execution feedback and design a test-driven discriminative task to train a novel discriminator, aiming to let the model distinguish whether the generated code is correct or not. The preliminary results on a newly published dataset demonstrate the effectiveness of our proposed framework for code generation. Particularly, in terms of the [email protected] metric, we achieve 8.81 and 11.53 gains compared with CodeGPT and CodeT5, respectively.

By analyzing users’ behavior data for personalized services, most state-of-the-art methods for user preference modeling are often based on batch-mode machine learning algorithms, where all rating data are assumed to be available throughout the training process. However, data in the real world often arrives sequentially and user preference may change dynamically. The real-time characteristics of rating data make the algorithms for preference modeling challenging to suit real-world online applications. By the user preference model (UPM) based on Bayesian network with a latent variable (BNLV), uncertain relationships among relevant attributes of users, objects and ratings could be represented, in which user preference is represented by the latent variable. In this paper, we propose an online approach for parameter learning of UPM. Specifically, we first extend the classic Voting EM algorithm by using Bayesian estimation in terms of the situation with latent variables. Consequently, we propose the algorithm for learning parameters of UPM from few and sequentially-changing rating data to reflect the gradually changing preferences. Finally, we test the effectiveness of our proposed algorithm by conducting experiments on various datasets. Experimental results demonstrate the superiority of our method in various measurements.

The popularity of convolutional architecture has made sensor-based human activity recognition (HAR) become one primary beneficiary. By simply superimposing multiple convolution layers, the local features can be effectively captured from multi-channel time series sensor data, which could output high-performance activity prediction results. On the other hand, recent years have witnessed great success of Transformer model, which uses powerful self-attention mechanism to handle long-range sequence modeling tasks, hence avoiding the shortcoming of local feature representations caused by convolutional neural networks (CNNs). In this paper, we seek to combine the merits of CNN and Transformer to model multi-channel time series sensor data, which might provide compelling recognition performance with fewer parameters and FLOPs based on lightweight wearable devices. To this end, we propose a new Dual-branch Interactive Network (DIN) that inherits the advantages from both CNN and Transformer to handle multi-channel time series for HAR. Specifically, the proposed framework utilizes two-stream architecture to disentangle local and global features by performing conv-embedding and patch-embedding, where a co-attention mechanism is used to adaptively fuse global-to-local and local-to-global feature representations. We perform extensive experiments on three mainstream HAR benchmark datasets including PAMAP2, WISDM, and OPPORTUNITY, which verify that our method consistently outperforms several state-of-the-art baselines, reaching an F1-score of 92.05%, 98.17%, and 91.55% respectively with fewer parameters and FLOPs. In addition, the practical execution time is validated on an embedded Raspberry Pi P3 system, which demonstrates that our approach is adequately efficient for real-time HAR implementations and deserves as a better alternative in ubiquitous HAR computing scenario. Our model code will be released soon.

Forecasting traffic inflows and outflows is crucial for intelligent transportation applications such as traffic management and risk assessment. Recently, deep learning models, which focus on capturing spatio-temporal correlations between stations (locations) by constructing Spatio-Temporal Feature Learners (STFL), have achieved promising performance in traffic inflows and outflows prediction. However, two unresolved issues limit the performance of these models. i) dynamic and heterogeneous intra- and inter-relationships between flows are ignored, and ii) the STFL in these models cannot capture the global information. To address the above issues, we propose a novel deep Spatio-Temporal Network framework based on Multi-Relational learning (MR-STN) for predicting traffic inflows and outflows. Specifically, a multi-relational learning module is designed to comprehensively model three kinds of relationships between flows while extracting diverse spatio-temporal features. In this module, an enhanced STFL is developed to capture both local and global information. Then, a feature fusion module is introduced to extract fused features for inflows and outflows respectively via a gated fusion mechanism. On this basis, the prediction module uses fusion features to generate future inflows and outflows. Finally, we implement the proposed framework with four state-of-the-art graph-based deep spatio-temporal models to demonstrate its generality and superiority. Extensive experiments on three datasets show that the proposed framework can significantly boost the performance of existing models.

With the booming of deep learning, massive attention has been paid to developing neural models for multilabel text categorization (MLTC). Most of the works concentrate on disclosing word-label relationship, while less attention is taken in exploiting global clues, particularly with the relationship of document-label. To address this limitation, we propose an effective collaborative representation learning (CRL) model in this article. CRL consists of a factorization component for generating shallow representations of documents and a neural component for deep text-encoding and classification. We have developed strategies for jointly training those two components, including an alternating-least-squares-based approach for factorizing the pointwise mutual information (PMI) matrix of label-document and multitask learning (MTL) strategy for the neural component. According to the experimental results on six data sets, CRL can explicitly take advantage of the relationship of document-label and achieve competitive classification performance in comparison with some state-of-the-art deep methods.

With the widespread deployment of various cloud computing and service-oriented systems, there is a rapidly increasing demand for collaborative quality-of-service (QoS) prediction. Existing QoS prediction methods have made great progress in modeling users and services as well as exploiting contexts of service invocations. However, they ignore the completion of service requests/responses relies on the underlying network topology and the complex interactions between Autonomous Systems. To tackle this challenge, we propose a topology-aware neural (TAN) model for collaborative QoS prediction. In the TAN model, the features of users, services, and intermediate nodes on the communication path are projected to a shared latent space as input features. To jointly characterize the invocation process, the path features and end-cross features are captured respectively through an explicit path modeling layer and an implicit cross-modeling layer. After that, a gating layer fuses and transmits these features to the prediction layer for estimating unknown QoS values. In this way, TAN provides a flexible framework that can comprehensively capture the invocation context for making accurate QoS prediction. Experimental results on two real-world datasets demonstrate that TAN significantly outperforms state-of-the-art methods on the tasks of response time, throughput, and reliability prediction. Also, TAN shows better extensibility of using auxiliary information.

As the number of Web APIs ever increases, choosing the appropriate APIs for mashup creations becomes more difficult. To tackle this problem, various methods have been proposed to recommend APIs to match requirements of mashups and achieved much success. However, there existed some challenges with feature fusion and utilization, textual requirement understanding, utilization of Mashup categories and compatibility evaluation. Therefore, we propose a neural framework (MTFM) based on multi-model fusion and multi-task learning for Mashup-oriented Web API recommendation. MTFM exploits a semantic component to generate representations of requirements and introduces a feature interaction component to model the feature interaction between mashups and Web APIs. Output features of both components are further fused to predict the candidate APIs, and this enables us to have both the advantages of content-based and collaborative filtering methods. We further introduce mashup category judgment as an auxiliary task, where both tasks are viewed as a multi-label learning problem and jointly optimized with multi-task learning. Also, we have extended MTFM to MTFM++ to take advantage of the metadata and quality features of APIs, and proposed a metric for compatibility evaluation. Experimental results on the ProgrammableWeb dataset show that our methods outperform most popular state-of-the-art methods.

POI (point-of-interest) recommendation as an important type of location-based services has received increasing attention with the rise of location-based social networks. Although significant efforts have been dedicated to learning and recommending users' next POIs based on their historical mobility traces, there still lacks consideration of the discrepancy of users' check-in time preferences and the inherent relationships between POIs and check-in times. To fill this gap, this paper proposes a novel recommendation method which applies multi-task learning over historical user mobility traces known to be sparse. Specifically, we design a cross-graph neural network to obtain time-aware user modeling and control how much information flows across different semantic spaces, which makes up the inadequate representation of existing user modeling methods. In addition, we design a check-in time prediction task to learn users' activities from a time perspective and learn internal patterns between POIs and their check-in times, aiming to reduce the search space to overcome the data sparsity problem. Comprehensive experiments on two real-world public datasets demonstrate that our proposed method outperforms several representative POI recommendation methods with 8.93% to 20.21 % improvement on Recall@1, 5, 10, and 9.25% to 17.56% improvement on Mean Reciprocal Rank.

Bundle recommendation plays a crucial role in the service ecosystem. However, most existing bundle recommendation methods are limited in several critical aspects such as the lack of injecting different relations into the representations of bundles and items, and the ignorance of neighbor interactions. To address these limitations, in this paper, we propose a relational graph neural network with neighbor interactions for bundle recommendation. Specifically, we firstly construct two relational graphs, e.g., user-bundle-item interaction graph and bundle-item affiliation graph. We utilize a relational graph neural network to inject different relations into representations of bundles and items. Secondly, we consider neighbor interactions to highlight common properties of neighbors. Finally, a multi-task learning framework is also exploited to capture users' preferences at the item level to further enhance bundle recommendation performance. Comprehensive experiments on two real-world public datasets demonstrate that our proposed method can outperform various representative bundle recommendation methods.

In recent years, various collaborative QoS prediction methods have been put forward to coping with the demand for efficient quality-of-service (QoS) evaluation, by drawing lessons from the recommender systems. However, there still remain some challenging issues on this direction, as how to effectively exploit complex contexts to improve prediction accuracy, and how to realize collaborative QoS prediction of multiple attributes. Inspired by the principles of deep learning, we have proposed a universal deep neural model (DNM) for making multiple attributes QoS prediction with contexts. In this model, contextual features are mapped into a shared latent space to semantically characterize them in the embedding layer. The contextual features with their higher-order interactions are captured through the interaction layer and the perception layers. Multi-tasks prediction is realized by stacking task-specific perception layers on the shared neural layers. Armed with these, DNM provides a powerful framework to integrate with various contextual features to realize multi-attributes QoS prediction. Experimental results from a large-scale QoS-specific dataset demonstrate that DNM achieves superior prediction accuracy in term of mean absolute error (MAE) compared with the state-of-the-art collaborative QoS prediction techniques. Additionally, the DNM model has a good robustness and extensibility on exploiting heterogeneous contextual features.

Automated service classification is the foundation for service discovery and service composition. Currently, many existing methods extracting features from functional description documents suffer the problem of data sparsity. However, beside functional description documents, the Web API ecosystem has accumulated a wealth of information that can be used to improve the accuracy of Web service (API) classification. At the moment, there is an absence of a unified way to combine functional description documents with other sources of information (e.g., attributes, interactions and external knowledge) accumulated in the Web API ecosystem for API classification. To address this issue, we present a dual-GCN framework that can effectively suppress the noise propagation of textual contents by distinguishing functional description documents and other sources of information (specifically Mashup-API co-invocation patterns by default in this paper) for API classification. This framework is extensible with the ability to include different sources of information accumulated in the Web API ecosystem. Comprehensive experiments on a real-world public dataset demonstrate that our proposed method can outperform various representative methods for API classification.

Self-Admitted Technical Debt (SATD) is a sub-type of technical debt. It is introduced to represent such technical debts that are intentionally introduced by developers in the process of software development. While being able to gain short-term benefits, the introduction of SATDs often requires to be paid back later with a higher cost, e.g., introducing bugs to the software or increasing the complexity of the software.

A key problem in online social networks is the identification of users' link information and the analysis of how these are reflected in the recommender systems. The basis to tackle this issue is user similarity measures. In this paper, we propose non-negative multiple matrix factorization with social similarity for recommender systems, considering the similarities between users, the relationships of users-resources and tags-resources. On this basis, we comparatively analyzed different performances of the recommendation with every similarity measure between users. In addition, our method can also recommend friends, resources, and tags to users. Experimental results on Lastfm and Delicious datasets show that the proposed method can significantly improve the recommendation accuracy compared with the art collaborative filtering methods.

Existing route recommendation systems have two main weaknesses. First, they usually recommend the same route for all users and cannot help control traffic jam. Second, they do not take full advantage of real-time traffic to recommend the best routes. To address these two problems, we develop a real-time route recommendation system, called R3, aiming to provide users with the real-time-traffic-aware routes. R3 recommends diverse routes for different users to alleviate the traffic pressure. R3 utilizes historical taxi driving data and real-time traffic data and integrates them together to provide users with real-time route recommendation.

Query term suggestion that interactively expands the queries is an indispensable technique to help users formulate high-quality queries and has attracted much attention in the community of web search. Existing methods usually suggest terms based on statistics in documents as well as query logs and external dictionaries, and they neglect the fact that the topic information is very crucial because it helps retrieve topically relevant documents. To give users gratification, we propose a novel term suggestion method: as the user types in queries letter by letter, we suggest the terms that are topically coherent with the query and could retrieve relevant documents instantly. For effectively suggesting highly relevant terms, we propose a generative model by incorporating the topical coherence of terms. The model learns the topics from the underlying documents based on Latent Dirichlet Allocation (LDA). For achieving the goal of instant query suggestion, we use a trie structure to index and access terms. We devise an efficient top-k algorithm to suggest terms as users type in queries. Experimental results show that our approach not only improves the effectiveness of term suggestion, but also achieves better efficiency and scalability.

Backfilling is well known in parallel job scheduling to increase system utilization and user satisfaction over traditional non-backfilling scheduling algorithms, which allow small jobs from the back of the queue to execute before larger jobs arriving earlier, and resources could be reserved to protect the latter from starvation. This paper proposed a relaxed backfill scheduling mechanism supporting multiple reservations, and investigated its effectiveness in reducing the average waiting time and average slowdown of jobs by using simulations with real traces. Different from existing relaxed scheduling, which restrict the maximum number of reservations to one, this new mechanism can support the relaxation of multiple reservations and works efficiently in scheduling by successful avoidance of raising chain reactions in relaxing the start times of multiple already existing reservations. Experimental results suggest that although the performances of both the relax-based backfilling and the strict backfill depend on the accuracy of runtime estimates, reservation depths, traces and system load alike, the former scheduling is more flexible and generally more effective in reducing the average waiting time and average slowdown of jobs, without loss of utilization.

Abstract In this article, we propose to apply the topic model and topic‐level eigenfactor (TEF) algorithm to assess the relative importance of academic entities including articles, authors, journals, and conferences. Scientific impact is measured by the biased PageRank score toward topics created by the latent topic model. The TEF metric considers the impact of an academic entity in multiple granular views as well as in a global view. Experiments on a computational linguistics corpus show that the method is a useful and promising measure to assess scientific impact.

Form-style interfaces have been widely used to allow users to access information. In this demonstration paper, we develop a new search paradigm in form-style query interfaces, called Seaform (which stands for Search-As-You-Type in Forms), which computes answers on-the-fly as a user types in a query letter by letter and gives the user instant feedback. Seaform provides better user experiences compared with traditional form-based query systems by reducing the efforts for a user to compose a high-quality query to find relevant answers. Seaform can also enhance faceted search and allow users to on-the-fly explore the underlying data. This search paradigm requires high performance to achieve an interactive speed. We develop efficient techniques and use them to implement two systems on real datasets. We demonstrate the features of these systems.