Session-based recommendation (SBR) aims to predict anonymous users' next interaction based on their interaction sessions. In practical recommendation scenario, low-exposure items constitute the majority of interactions, creating a long-tail distribution that severely compromises recommendation diversity. Existing approaches attempt to address this issue by promoting tail items but incur accuracy degradation, exhibiting a "see-saw" effect between long-tail and accuracy performance. We attribute such conflict to session-irrelevant noise within the tail item set, which existing long-tail approaches fail to identify and constrain effectively. To resolve our fundamental conflict, we propose HID (Hybrid Intent-based Dual Constraint Framework), a plug-and-play framework that transforms the conventional "see-saw" into a "win-win" relationship through introducing the hybrid intent-based dual constraints. Two key innovations are incorporated in this framework: (i) Hybrid Intent Learning, where we reformulate the intent extraction strategies by employing attribute-aware spectral clustering to reconstruct the item-to-intent mapping. Furthermore, discrimination of session-irrelevant noise is achieved through the assignment of both target and noise intents to each sessions. (ii) Intent Constraint Loss, where we propose two novel constraint paradigms regarding the diversity and accuracy to regulate the representation learning process, and unify the two optimization objectives into a unique loss. Extensive experiments across multiple SBR models and datasets demonstrate that HID can enhance both long-tail performance and recommendation accuracy, establishing new state-of-the-art performance in long-tail recommender systems.

Event analysis is an essential and fundamental direction of information extraction, involving various event-centric tasks at different granularity of documents. While large language models (LLMs) have preliminarily achieved promising performance in part of these tasks individually, their capability in event analysis still lacks comprehensive understanding due to restricted document granularity, task designs, and data source of existing benchmarks. To address these limitations, we introduce MiGUE-Bench, a systematic benchmark for assessing the performance of LLMs in multi-granularity event analysis. To support large-scale evaluation, we first develop an LLM-driven self-correcting annotation framework called MiGUE-Pipeline, enabling scalable acquisition of high-quality source data of events with automatic labels. Then, we design four core tasks in our benchmark, i.e., event detection, relation reasoning, structure induction, and future prediction, to probe model competence at different levels, from atomic event details to complex cross-document narratives. Extensive experiments on state-of-the-art LLMs and retrieval-augmented generation (RAG) methods delineate the current capability boundary and identify critical deficiencies, providing insights into the future improvement of LLMs in challenging event analysis tasks.

Large language models (LLMs) show promising performance on small-scale graph reasoning tasks but fail when handling real-world graphs with complex queries. This phenomenon arises from LLMs' working memory constraints, which result in their inability to retain long-range graph topology over extended contexts while sustaining coherent multi-step reasoning. However, real-world graphs are often structurally complex, such as Web, Transportation, Social, and Citation networks. To address these limitations, we propose GraphCogent, a collaborative agent framework inspired by human Working Memory Model that decomposes graph reasoning into specialized cognitive processes: sense, buffer, and execute. The framework consists of three modules: Sensory Module standardizes diverse graph text representations via subgraph sampling, Buffer Module integrates and indexes graph data across multiple formats, and Execution Module combines tool calling and tool creation for efficient reasoning. We also introduce Graph4real, a comprehensive benchmark that contains four domains of real-world graphs (Web, Transportation, Social, and Citation) to evaluate LLMs' graph reasoning capabilities. Our Graph4real covers 21 different graph reasoning tasks, categorized into three types (Structural Querying, Algorithmic Reasoning, and Predictive Modeling tasks), with graph scales up to 10 times larger than existing benchmarks. Experiments show that Llama3.1-8B based GraphCogent achieves a 50% improvement over massive-scale LLMs like DeepSeek-R1 (671B). Compared to state-of-the-art code-based baseline, our framework outperforms by 20% in accuracy while reducing token usage by 80% for in-toolset tasks and 30% for out-toolset tasks.

Foodborne diseases (FBDs) pose a significant threat to public health worldwide, with clustered outbreaks resulting in substantial economic and healthcare burdens. Conventional surveillance methods struggle to capture the multi-dimensional, complex relationships among cases, limiting early warning and targeted intervention. To address this gap, we propose a data-driven framework for FBD outbreak risk prediction by integrating patient association graphs and an improved graph attention network (GAT). First, we construct the patient association graph, where nodes represent individual FBD cases and edge weights are derived from the fusion of four key similarity dimensions, including spatiotemporal similarity (STS), food exposure similarity (FES), symptom similarity (SS), and demographic similarity (DS). Second, we introduce GatedResGAT, a novel graph attention network (GAT) enhanced with gated residual connections (GRC) and skip residual connections (SRC), to alleviate feature over-smoothing and gradient vanishing. Evaluated on real-world surveillance data from Wuxi, China, GatedResGAT achieves the area under the precision-recall curve (PR-AUC) of 0.7553, F1-Score of 0.7465, and Matthews Correlation Coefficient (MCC) of 0.7050, outperforming the best baseline model by 7.12%, 6.86%, and 12.64%, respectively. Ablation studies verify that FES and STS are the primary indicators for identifying case clustering, and both residual connections significantly enhance model performance. Early warning analysis of FBD outbreaks identifies high-risk transmission sub-networks with central outbreak nodes and potential contamination sources. These findings provide public health authorities with direct evidence for rapid source tracing and precise intervention

Dataset distillation (DD) has demonstrated the promise of synthesizing smaller datasets that enable competitive performance. While most of DD methods operate in pixel space, they often suffer from poor scalability on high-resolution datasets. Recent works have thus shifted towards parameterizing synthetic data using deep generative priors. However, these approaches apply latent updates to all generator layers, leading to substantial computational overhead. To address this limitation, we propose Generative Lightweight Distillation (GLiD), a unified framework that jointly compresses the generator and accelerates latent optimization for efficient distillation. GLiD introduces two key components: (1) We introduce a Classification-Diversity Sensitivity Pruning mechanism that quantifies both discriminative utility and semantic diversity of each output channel to guide structural pruning and selective layer-wise optimization. (2) We also present a Layer-Adaptive Scheduling strategy that dynamically allocates latent update steps across generator stages based on convergence behavior. Extensive experiments on CIFAR-10 and ImageNet-1K and its subsets demonstrate that our GLiD achieves up to 10× acceleration across different datasets, while maintaining performance competitive with state-of-the-art methods.

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.

Training-based algorithms significantly outperform training-free methods in terms of recognition performance for steady-state visual-evoked potential (SSVEP)-based brain-computer Interfaces (BCIs). However, collecting training data requires calibration experiments that are effort-intensive and often costly. These calibration demands limit the practicality of BCI, as users (and even system operators) may experience fatigue or lose interest in continued use. Transfer learning (TL) offers an effective solution, but it typically relies on either a certain amount of target domain data or extensive source domain data. To address this limitation, we introduce the concept of cross-dataset TL in SSVEP for the first time to extract transfer knowledge from other datasets. During this process, we identified a data mismatch problem that severely compromises the generalizability of transfer knowledge. To overcome this challenge, we propose a TL-SSVEP decoding algorithm calibrated with single-trial data (TL-CSTD). Specifically, we use 2 s of 8 Hz single-trial calibration data from the target domain to obtain matched transfer templates from the source domain. These templates are then corrected to extract holistic and single-period transfer knowledge, which are subsequently employed to construct an efficient TL-SSVEP decoding model for the target subject. Experimental results on three large SSVEP datasets demonstrate that TL-CSTD effectively addresses the data mismatch problem and achieves excellent SSVEP recognition performance using only 2 s of single-trial calibration data, showing its significant application potential and practicality.

Session-based recommendation (SBR) predicts the next interaction of anonymous users based on their session data. In practical scenarios, low-exposure tail items constitute the majority of interactions, leading to the biased recommendation whereby a small set of high-exposure head items is preferentially recommended. Existing solutions focus on promoting those tail items but with a sacrifice in recommendation accuracy. We attribute such a tradeoff to session-irrelevant noise within the tail items, which existing long-tail approaches fail to identify and constrain effectively. To resolve this fundamental conflict, we propose Hybrid Intent-based Dual Constraint framework (HID) , a plug-and-play framework that converts the traditional tradeoff between long-tail performance and accuracy into a synergistic relationship by introducing hybrid intent-based dual constraints. HID incorporates two key innovations: (i) Hybrid Intent Learning , which reformulates intent extraction through attribute-aware spectral clustering and discriminates session-irrelevant noise by assigning target and noise intents to each session and (ii) Intent Constraint Loss , which introduces two constraints—diversity and accuracy—to regulate item and session representation learning. These two objectives are unified into a single training loss through rigorous theoretical derivation. To handle ambiguous user intent in the real-world, we also propose a slack version of HID. Extensive experiments show that HID enhances both long-tail performance and accuracy, establishing new state-of-the-art performance in long-tail recommender systems. The implementation code is available at: https://github.com/jarviswww/Code4TradeSynergy .

Contemporary deep learning, characterized by the training of cumbersome neural networks on massive datasets, confronts substantial computational hurdles. To alleviate heavy data storage burdens on limited hardware resources, numerous dataset compression methods such as dataset distillation (DD) and coreset selection have emerged to obtain a compact but informative dataset through synthesis or selection for efficient training. However, DD involves an expensive optimization procedure and exhibits limited generalization across unseen architectures, while coreset selection is limited by its low data keep ratio and reliance on heuristics, hindering its practicality and feasibility. To address these limitations, we introduce a newly versatile framework for dataset compression, namely Adaptive Dataset Quantization (ADQ). Specifically, we first identify the sub-optimal performance of naive Dataset Quantization (DQ), which relies on uniform sampling and overlooks the varying importance of each generated bin. Subsequently, we propose a novel adaptive sampling strategy through the evaluation of generated bins' representativeness score, diversity score and importance score, where the former two scores are quantified by the texture level and contrastive learning-based techniques, respectively. Extensive experiments demonstrate that our method not only exhibits superior generalization capability across different architectures, but also attains state-of-the-art results.

Multimodal dialogue generation requires a comprehensive understanding of each modality and the ability to effectively integrate these modalities to produce contextually relevant and diverse responses. However, existing approaches, which primarily rely on language models, often suffer from modality bias, stemming from an over-dependence on language data in training, and inherent biases in model structure. This over-reliance results in an imbalance across modalities and leads to suboptimal performance. To address this challenge, we propose a novel dialogue generation model with Knowledge-enhanced Multimodal Graph, dubbed KMG-LL, designed to balance multimodal information and incorporate external knowledge. Specifically, we transform multimodal information into graph-structured data and integrate them with commonsense knowledge to construct knowledge-enhanced multimodal graph. To further refine the extraction of multimodal information, we propose a modal-balanced knowledge aggregation that processes modalities cross multiple levels. Extensive experiments on two multimodal dialogue datasets demonstrate that KMG-LL significantly outperforms existing baselines in multimodal dialogue generation.

Audio-to-intent recognition is a critical task focused on identifying a speaker’s intent from spoken language. Recently, multimodal audio-to-intent approaches have emerged as the predominant strategy for enhancing audio-based intent recognition. In this study, we conduct a series of empirical experiments that reveal a significant modality bias in current multimodal audio-to-intent recognition methods. Specifically, these methods disproportionately rely on the textual modality to determine intent, often neglecting the audio data. To address this issue, we propose a context-enhanced contrastive learning framework designed to capture rich regional- and global-audio context information, thereby enabling more balanced audio-to-intent recognition. Additionally, we introduce a prototype-based intent classification strategy that encourages different intent classes and modalities to converge toward unified prototypes, leading to smoother classification boundaries as opposed to the traditionally skewed boundaries. Extensive experiments demonstrate that our approach effectively mitigates modality bias, e.g., a performance improvement of 2.12% in intent classification compared to the state-of-the-art method GZAIR on the dataset MintRec.

Panoptic Scene Graph Generation (PSG) integrates instance segmentation with relation understanding to capture pixel-level structural relationships in complex scenes. Although recent approaches leveraging pre-trained vision-language models (VLMs) have significantly improved performance in the open-vocabulary setting, they commonly ignore the inherent limitations of VLMs in spatial relation reasoning, such as difficulty in distinguishing object relative positions, which results in suboptimal relation prediction. Motivated by the denoising diffusion model's inversion process in preserving the spatial structure of input images, we propose SPADE (SPatial-Aware Denoising-nEtwork) framework -- a novel approach for open-vocabulary PSG. SPADE consists of two key steps: (1) inversion-guided calibration for the UNet adaptation, and (2) spatial-aware context reasoning. In the first step, we calibrate a general pre-trained teacher diffusion model into a PSG-specific denoising network with cross-attention maps derived during inversion through a lightweight LoRA-based fine-tuning strategy. In the second step, we develop a spatial-aware relation graph transformer that captures both local and long-range contextual information, facilitating the generation of high-quality relation queries. Extensive experiments on benchmark PSG and Visual Genome datasets demonstrate that SPADE outperforms state-of-the-art methods in both closed- and open-set scenarios, particularly for spatial relationship prediction.

Large language models (LLMs) have been demonstrated to possess the capabilities to understand fundamental graph properties and address various graph reasoning tasks. Existing methods fine-tune LLMs to understand and execute graph reasoning tasks by specially designed task instructions. However, these Text-Instruction methods generally exhibit poor performance. Inspired by tool learning, researchers propose Tool-Instruction methods to solve various graph problems by special tool calling (e.g., function, API and model), achieving significant improvements in graph reasoning tasks. Nevertheless, current Tool-Instruction approaches focus on the tool information and ignore the graph structure information, which leads to significantly inferior performance on small-scale LLMs (less than 8B). To tackle this issue, we propose GraphTool-Instruction, an innovative Instruction-tuning approach that decomposes the graph reasoning task into three distinct subtasks (i.e., graph extraction, tool name identification and tool parameter extraction), and design specialized instructions for each subtask. Our GraphTool-Instruction can be used as a plug-and-play prompt for different LLMs without fine-tuning. Moreover, building on GraphTool-Instruction, we develop GTools, a dataset that includes twenty graph reasoning tasks, and create a graph reasoning LLM called GraphForge based on Llama3-8B. We conduct extensive experiments on twenty graph reasoning tasks with different graph types (e.g., graph size or graph direction), and we find that GraphTool-Instruction achieves SOTA compared to Text-Instruction and Tool-Instruction methods. Fine-tuned on GTools, GraphForge gets further improvement of over 30% compared to the Tool-Instruction enhanced GPT-3.5-turbo, and it performs comparably to the high-cost GPT-4o. Our codes and data are available at https://github.com/RongzhengWang/GraphTool-Instruction.

Recently, prompt tuning has garnered considerable attention due to its success across various Vision-Language (VL) tasks. However, unimodal prompts, coupled prompts, and joint prompts in these models often lead to suboptimal performance due to differences in information density and complexity between modalities. Particularly, in scenarios with missing modalities, these prompt-based approaches tend to exacerbate 'Channel Bias'-a phonomenon where models overly rely on specific feature (such as unmissing-modal feature) channels from the base tasks, thereby undermining the model's ability to capture crucial shared knowledge applicable to new tasks and affecting its generalizability. To address this challenge, we propose RobustPT, a dynamic disentanglement prompt tuning model designed to enhance the robustness of VL models under modality missing conditions. RobustPT utilizes a multi-channel prompting mechanism to dynamically disentangle and align prompts. Specifically, RobustPT is divided into single-channel tuning and alignment-channel tuning, where prompts for each modality run independently in sequence to delve deeply into their intrinsic characteristics, followed by an integration through a non-strong coupling strategy to effectively balance information contributions and enhance overall performance. Extensive experiments demonstrate that our RobustPT achieve significant improvements over the current state-of-the-art across all benchmark datasets. Our codes are available at https://github.com/Trae1ounG/RobustPT.

The growing demand for efficient deep learning has positioned dataset distillation as a pivotal technique for compressing training dataset while preserving model performance. However, existing inner-loop optimization methods for dataset distillation typically rely on random truncation strategies, which lack flexibility and often yield suboptimal results. In this work, we observe that neural networks exhibit distinct learning dynamics across different training stages—early, middle, and late—making random truncation ineffective. To address this limitation, we propose Automatic Truncated Backpropagation Through Time (AT-BPTT), a novel framework that dynamically adapts both truncation positions and window sizes according to intrinsic gradient behavior. AT-BPTT introduces three key components: (1) a probabilistic mechanism for stage-aware timestep selection, (2) an adaptive window sizing strategy based on gradient variation, and (3) a low-rank Hessian approximation to reduce computational overhead. Extensive experiments on CIFAR-10, CIFAR-100, Tiny-ImageNet, and ImageNet-1K show that AT-BPTT achieves state-of-the-art performance, improving accuracy by an average of 6.16\% over baseline methods. Moreover, our approach accelerates inner-loop optimization by 3.9 × while saving 63\% memory cost.

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.

Abstract Session-based recommendation (SBR) focuses on predicting the next potential item for anonymous users based on short-click sessions. However, these interaction sessions often contain noise items, which arise from misclicks or shifts in user interests. Existing denoising methods typically presume a strong exclusionary relationship between noise items and the recommendation target, assuming that reducing noise can enhance recommendation accuracy. In contrast, our observations reveal a nuanced phenomenon: as the length of the interaction session shortens, the effect of noise removal on recommendation performance gradually transitions from positive to negative. This finding suggests that in short sessions with insufficient contextual information, relying solely on the exclusion of noise items within the session may fail to improve and could even hinder-the recommendation performance. Such complexities have been largely overlooked in prior research. To bridge this gap, we propose two solutions: (i) expanding the view of denoising from a single session to multiple sessions (i.e., from local to global), and (ii) introducing relevant contextual information into each session by employing enhancement strategies. Therefore, we design the Hybrid Prototype-based In-and-Out Flow Network (HyPro), which employs both denoising and enhancing processes for each session based on our proposed hybrid prototypes. Specifically, for each item, HyPro first learns the hybrid prototype by aggregating information from the item’s semantic and topological neighbors across all sessions. Then, based on the hybrid prototypes, HyPro employs an in-and-out flow network comprising two components: (i) the out-flow channel, which targets the removal of irrelevant information at both the data and feature levels, and (ii) the in-flow channel, which integrates global information for each session at the item and session levels. Extensive experiments conducted on three real-world datasets demonstrate that HyPro outperforms the state-of-the-art baselines. The implementation code is available at: https://github.com/jarviswww/Code4HyPro .

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.

Multimodal knowledge graphs (MMKGs) have gained widespread adoption across various domains. However, existing transformer-based methods for MMKG representation learning primarily focus on enhancing representation performance, while overlooking time and memory costs, which reduces model efficiency. To tackle these limitations, we introduce a multimodal lightweight transformer (MLFormer) model, which not only ensures robust representation capabilities but also considerably improves computational efficiency. We find that the self-attention mechanism in transformers leads to substantial performance overheads. As a result, we optimize the traditional MMKGE model in two aspects: modality processing and modality fusion, by incorporating a filter gate and Fourier transform. Our experimental results on real-world multimodal knowledge graph completion datasets demonstrate that MLFormer achieves significant improvements in computational efficiency while maintaining competitive performance.

Prompt reasoning methods often suffer from insufficient facts or factual errors when dealing with commonsense reasoning tasks. And in multi-step reasoning methods such as Chain-of-Thought, these errors can accumulate to the extent of yielding incorrect answers. Recent advancements in Tree-of-Thought reasoning aim to tackle the issue of error accumulation. Nevertheless, applying this approach to commonsense reasoning tasks proves challenging because these tasks resist easy decomposition in a standardized manner.We proposed ToFC, a prompt-based reasoning method for solving above problems in commonsense reasoning tasks. ToFC is a step-by-step, tree-like framework for prompt reasoning. It accomplishes stepwise reasoning and fact enhancement through continued Fact-Proposals; heuristically plans reasoning paths through Self-Evaluation; and balances accuracy and efficiency with a modified Best-First Search algorithm. ToFC tackles the issue of error accumulation in current prompt-based commonsense reasoning methods and addresses inefficiencies in the Tree-of-Thought framework. Experimental results indicate that ToFC outperforms existing prompt reasoning methods, and its extra overhead is within a reasonable range.

In recent years, thanks to the vigorous development of deep learning, single image super-resolution has advanced greatly. Most super-resolution (SR) methods use convolution layers to construct the network, which achieves superior results over the traditional methods based on manual features. However, most methods based on convolutional neural networks (CNN) blindly deepen the depth of the network leading to a large number of model parameters, which inevitably brings huge computing overhead and memory consumption, and greatly limits the application in resource-limited devices. In order to alleviate this problem, a knowledge distillation framework based on contrastive learning is proposed to compress and accelerate the SR model with enormous parameters. The student network is directly constructed by reducing the number of layers of the teacher network. In particular, the proposed method distills the statistical information of the intermediate feature maps from the teacher network to train the lightweight student network. In addition, through explicit knowledge transfer, a novel contrastive loss is introduced to improve the reconstruction performance of the student network. Experiments show that the proposed contrastive distillation framework can effectively compress the model scale with an acceptable loss of performance.

Data-free knowledge distillation (DFKD) has emerged as a pivotal technique in the domain of model compression, substantially reducing the dependency on the original training data. Nonetheless, conventional DFKD methods that employ synthesized training data are prone to the limitations of inadequate diversity and discrepancies in distribution between the synthesized and original datasets. To address these challenges, this paper introduces an innovative approach to DFKD through diverse diffusion augmentation (DDA). Specifically, we revise the paradigm of common data synthesis in DFKD to a composite process through leveraging diffusion models subsequent to data synthesis for self-supervised augmentation, which generates a spectrum of data samples with similar distributions while retaining controlled variations. Furthermore, to mitigate excessive deviation in the embedding space, we introduce an image filtering technique grounded in cosine similarity to maintain fidelity during the knowledge distillation process. Comprehensive experiments conducted on CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets showcase the superior performance of our method across various teacher-student network configurations, outperforming the contemporary state-of-the-art DFKD methods. Code will be available at: https://github.com/SLGSP/DDA.

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.

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.

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

The Chebyshev map is a typical chaotic map which has been widely investigated for cryptography. However, researchers argue that almost all the encryption algorithms based on some chaotic maps are not as secure as they are announced. In this paper, the limitations of some multicast key agreement schemes are discussed. To eliminate those false and remain novelty, we consider how the chaotic maps can be improved to fulfill the cryptographic demands. To achieve this, we extend the domain of Chebyshev maps from real number set to finite field. Based on this extension, this paper proposes a novel multicast key exchange algorithm. By detailed experiments and analysis, we show that this algorithm is secure and efficient. As far as we know, such a multicast key exchange algorithm has not been reported. The results given here, to the best of our knowledge, are novel.

The Adachi neural network (AdNN) is a fascinating neural network (NN) which has been shown to possess chaotic properties, and to also demonstrate associative memory (AM) and pattern recognition (PR) characteristics. Variants of the AdNN have also been used to obtain other PR phenomena, and even blurring. An unsurmountable problem associated with the AdNN and the variants referred to above is that all of them require a quadratic number of computations. This is essentially because the NNs in each case are completely connected graphs. In this paper, we consider how the computations can be significantly reduced by merely using a linear number of computations. To achieves this, we extract from the original completely connected graph one of its spanning trees. We then address the problem of computing the weights for this spanning tree. This is done in such a manner that the modified tree-based NN has approximately the same input-output characteristics, and thus the new weights are themselves calculated using a gradient-based algorithm. By a detailed experimental analysis, we show that the new linear-time AdNN-like network possesses chaotic and PR properties for different settings. As far as we know, such a tree-based AdNN has not been reported, and the results given here are novel.

In the past decades, distributed systems have been widely applied to real-time applications, most of which have fault-tolerance requirements to assure high reliability. Due to the stringent space constraints of real-time systems, the issue of schedulability becomes a major concern in the design of fault-tolerant and real-time distributed systems. Most existing real-time and fault-tolerant scheduling algorithms, which are based on the primary-backup scheme for periodic real-time tasks, introduce unnecessary redundancies by aggressively using active-backup copies. To solve this problem, we propose two novel fault-tolerant techniques, which are seamlessly integrated with fixed-priority-based scheduling algorithms. These techniques leverage redundancies to enhance schedulability in fault-tolerant and real-time distributed systems. Our fault-tolerant techniques make use of the primary-backup scheme to tolerate permanent hardware failures. The first technique (referred to as Tercos) terminates the execution of active-backup copies, when corresponding primary copies are successfully completed. Tercos is designed to reduce scheduling lengths in fault-free scenarios to enhance schedulability by virtue of executing portions of active-backup copies in passive forms. The second technique (referred to as Debus) uses a deferred-active-backup scheme to further minimize schedule lengths to improve the schedulability performance. Debus schedules active-backup copies as late as possible, while terminating active-backup copies when their primary copies are completed. Experimental results show that, compared with existing algorithms in literature, Tercos can significantly improve schedulability by up to 17.0% (with an average of 9.7%). Furthermore, empirical results reveal that Debus can enhance schedulability over Tercos by up to 12% (with an average of 7.8%).