arXiv subjects

Zhishu Shen

Publications and source records attributed to Zhishu Shen.

At least 19 recordsLinked to original sources

arXiv2026

Discovering and Preserving Category Correlation Knowledge via Adaptive Reciprocal Knowledge Distillation

Knowledge distillation aims to improve the performance of lightweight student models by transferring knowledge from larger and more powerful teacher models. However, a substantial size gap between teacher and student models often impedes effective knowledge transfer. Most existing approaches adopt a static, one-way teacher-to-student distillation paradigm, which overlooks the dynamic nature of student learning and fails to provide targeted guidance on hard samples. In this paper, we propose adaptive reciprocal knowledge distillation (AR-KD), a novel method that improves knowledge transfer by simplifying the teacher's output distribution. Specifically, AR-KD performs reciprocal adaptation on the teacher by matching its class correlation matrix to the student's relational representation, which reshapes the teacher's prediction structure to better suit the student's capacity. This relational alignment mitigates the collapse of inter-class dark knowledge caused by overconfident teachers, enabling the student to learn from richer and more compatible supervisory signals. We evaluate AR-KD on CIFAR-100 and ImageNet-1k classification datasets, where it outperforms state-of-the-art knowledge distillation baselines. Specifically, AR-KD improves student performance across homogeneous and heterogeneous setups: up to 7.13% accuracy gain for students, 1.42% to 4.15% higher than vanilla KD on average, and further improvements when integrated with other advanced methods. Our code is available at https://anonymous.4open.science/r/ARKD/.

cs.LG↗

arXiv2026

Counterfactual Benchmarking and Training for Factuality Consistency and Order-Robust Grounded Reasoning in LLMs over Heterogeneous Knowledge

Large language models (LLMs) have increasingly supported response generation grounded in user-provided knowledge spanning heterogeneous structures. However, existing benchmarks provide limited assessment of whether LLMs can faithfully perform multi-hop reasoning chains across such knowledge contexts while remaining robust to variations in their input order. We introduce TKFQA, a factuality consistency benchmark comprising 10,130 question-answering (QA) pairs grounded in tables, texts, and knowledge graphs (KGs). Each example is constructed from an explicit counterfactual reasoning chain, enabling the joint evaluation of answer correctness, reasoning-chain accuracy, and robustness to different input-order. An extensive evaluation of 14 open- and closed-source LLMs reveals that state-of-the-art models exhibit limited reasoning-chain accuracy and remain sensitive to variations in the input order of heterogeneous knowledge contexts. To address these limitations, we propose ORLF, an LLM-agnostic training framework that models cross-context topological relations through knowledge-specific latent vectors. ORLF integrates context-wise position encoding, a latent-bridge attention mask, and topological knowledge bias to preserve knowledge-specific bias and encode topological semantics. Experiments across four LLM backbones show that ORLF outperforms competitive training-free and LoRA-based baselines, improving average Exact Match and Reasoning-Chain Accuracy by 2.15% and 4.29%, respectively, while reducing order-induced performance standard deviation by 0.04% to 3.01%.

cs.AI↗

arXiv2026

Farm-LightSeek: An Edge-centric Multimodal Agricultural IoT Data Analytics Framework with Lightweight LLMs

Amid the challenges posed by global population growth and climate change, traditional agricultural Internet of Things (IoT) systems is currently undergoing a significant digital transformation to facilitate efficient big data processing. While smart agriculture utilizes artificial intelligence (AI) technologies to enable precise control, it still encounters significant challenges, including excessive reliance on agricultural expert knowledge, difficulties in fusing multimodal data, poor adaptability to dynamic environments, and bottlenecks in real-time decision-making at the edge. Large language models (LLMs), with their exceptional capabilities in knowledge acquisition and semantic understanding, provide a promising solution to address these challenges. To this end, we propose Farm-LightSeek, an edge-centric multimodal agricultural IoT data analytics framework that integrates LLMs with edge computing. This framework collects real-time farmland multi-source data (images, weather, geographic information) via sensors, performs cross-modal reasoning and disease detection at edge nodes, conducts low-latency management decisions, and enables cloud collaboration for model updates. The main innovations of Farm-LightSeek include: (1) an agricultural "perception-decision-action" closed-loop architecture; (2) cross-modal adaptive monitoring; and (3)a lightweight LLM deployment strategy balancing performance and efficiency. Experiments conducted on two real-world datasets demonstrate that Farm-LightSeek consistently achieves reliable performance in mission-critical tasks, even under the limitations of edge computing resources. This work advances intelligent real-time agricultural solutions and highlights the potential for deeper integration of agricultural IoT with LLMs.

cs.CV↗

arXiv2026

A RAG-Enhanced Bi-Level Cognitive Orchestration Framework for LEO Satellite Networks

The rapid growth of remote sensing data in Low Earth Orbit (LEO) satellite networks is increasingly constrained by limited downlink capacity to terrestrial networks. Satellite edge computing alleviates this pressure by enabling in-orbit data processing. However, it introduces a new challenge of spatio-temporal resource fragmentation. Variations in onboard computing capability, constrained energy availability, and intermittent inter-satellite and satellite-ground connectivity lead to highly dynamic and uneven resource distribution, which degrades the performance of conventional static routing and scheduling approaches. To address this, we propose a Retrieval-Augmented Generation (RAG)-enhanced bi-level cognitive orchestration framework for knowledge-guided, multi-objective scheduling. The proposed framework explicitly decouples network control across two different operational scales: at the strategic upper level, a Large Language Model (LLM) leverages an offline-distilled Expert Knowledge Base (EKB) to dynamically infer preference weights based on a compact abstract-state descriptor of real-time network conditions. At the lower execution level, a fidelity-aware genetic scheduler utilizes these inferred weights to compute physically feasible, collision-free joint routing and task offloading schedules. Extensive evaluations on a high-fidelity Walker-Delta network testbed under mixed-criticality workloads demonstrate that the proposed framework effectively consolidates fragmented resources, achieving a 30.7% reduction in packet loss, a 30% improvement in energy efficiency over the most competitive learning-based baseline, and an 8.5% decrease in end-to-end latency, while maintaining robust performance under cascading node-failure scenarios.

cs.DC↗

arXiv2026

Meta-Modal Agent: Sequential Evidence Routing for Missing-Modality Candidate Reranking

Missing modalities cause severe failures in multimodal recommender systems. User histories, item text, and visual evidence are frequently absent during cold-start scenarios, exactly when recommendation quality matters most. Existing approaches recover absent signals through imputation, feature propagation, or generative reconstruction, but these strategies can inject unsupported evidence when the surviving signals are weak. We introduce the Meta-Modal Agent (MMA), a large language model based candidate-pool reranker that treats missingness as a sequential evidence-routing problem. MMA is trained with balanced missingness-task reinforcement learning over masked-modality episodes and is evaluated in two variants: MMA-Auto, which uses only automated text, image, and graph tools, and MMA-Interactive, which additionally permits clarification questions grounded in surviving modalities as an upper-bound diagnostic. MMA operates after a first-stage retriever has produced a candidate pool; it scores those candidates rather than retrieving items from the full catalog. Final reranking fuses MMA scores with first-stage retrieval scores selected on validation data. Our evaluation is organized around four evidence checks required for a robust missing-modality claim: oracle-free one-observed-modality availability (OOMA) robustness, per-modality OOMA breakdowns, fixed-pool full-catalog reranking, and a deterministic-router mechanism control. MMA-Auto improves target-positive OOMA NDCG@10 by 4.0% and fixed-pool full-catalog reranking NDCG@10 by 12.7% over the strongest non-interactive baseline. RuleRouter-Fuse, which uses the same tools and fusion rule without learned policy updates, underperforms MMA-Auto, supporting learned routing beyond deterministic tool fusion. MMA-Interactive adds a 4.1% upper-bound gain when clarification is available.

cs.IR↗

arXiv2026

ML-ECS: A Collaborative Multimodal Learning Framework for Edge-Cloud Synergies

Edge-cloud synergies provide a promising paradigm for privacy-preserving deployment of foundation models, where lightweight on-device models adapt to domain-specific data and cloud-hosted models coordinate knowledge sharing. However, in real-world edge environments, collaborative multimodal learning is challenged by modality heterogeneity (different modality combinations across domains) and model-structure heterogeneity (different modality-specific encoders/fusion modules. To address these issues, we propose ML-ECS, a collaborative multimodal learning framework that enables joint training between a server-based model and heterogeneous edge models. This framework consists of four components: (1) cross-modal contrastive learning (CCL) to align modality representations in a shared latent space, (2) adaptive multimodal tuning (AMT) to preserve domain-specific knowledge from local datasets, (3) modality-aware model aggregation (MMA) to robustly aggregate while mitigating noise caused by missing modalities, and (4) SLM-enhanced CCL (SE-CCL) to facilitate bidirectional knowledge transfer between cloud and edge. Experimental results on various multimodal tasks show that \pname consistently outperform state-of-the-art baselines under varying modality availability, achieving improvements of 5.44% to 12.08% in Rouge-LSum and improving both client- and server-side performance. In addition, by communicating only low-rank LoRA parameters and fused representations, ML-ECS achieves high communication efficiency, requiring only 0.65% of the total parameter volume.

cs.DC↗

arXiv2026

Do We Really Need SFT? Prompt-as-Policy over Knowledge Graphs for Cold-start Next POI Recommendation

Next point-of-interest (POI) recommendation is a key component of smart urban services, yet it remains challenging under cold-start conditions with sparse user-POI interactions. Recent LLM-based methods address this issue through either supervised fine-tuning (SFT) or in-context learning (ICL), but SFT is costly and prone to overfitting active users, while static prompts in ICL lack adaptability to diverse user contexts. We argue that the main limitation lies not in LLM reasoning ability, but in how contextual evidence is constructed and presented. Accordingly, we propose Prompt-as-Policy over knowledge graphs (KG), a reinforcement-guided prompting framework that formulates prompt construction as a learnable decision process, while keeping the LLM frozen as a reasoning engine. To enable structured prompt optimization, we organize heterogeneous user-POI signals into a KG and transform mined relational paths into evidence cards, which serve as atomic semantic units for prompt composition. A contextual bandit learner then optimizes a prompt policy that adaptively determines (i) which relational evidences to include, (ii) how many evidences to retain per candidate POI, and (iii) how to organize and order them within the prompt. Experiments on three real-world datasets show that Prompt-as-Policy consistently outperforms state-of-the-art baselines, achieving an average 11.87% relative improvement in Acc@1 for inactive users, while maintaining competitive performance for active users, without any model fine-tuning.

cs.SI↗

arXiv2025

A Structure-Agnostic Co-Tuning Framework for LLMs and SLMs in Cloud-Edge Systems

The surge in intelligent applications driven by large language models (LLMs) has made it increasingly difficult for bandwidth-limited cloud servers to process extensive LLM workloads in real time without compromising user data privacy. To solve these problems, recent research has focused on constructing cloud-edge consortia that integrate server-based LLM with small language models (SLMs) on mobile edge devices. Furthermore, designing collaborative training mechanisms within such consortia to enhance inference performance has emerged as a promising research direction. However, the cross-domain deployment of SLMs, coupled with structural heterogeneity in SLMs architectures, poses significant challenges to enhancing model performance. To this end, we propose Co-PLMs, a novel co-tuning framework for collaborative training of large and small language models, which integrates the process of structure-agnostic mutual learning to realize knowledge exchange between the heterogeneous language models. This framework employs distilled proxy models (DPMs) as bridges to enable collaborative training between the heterogeneous server-based LLM and on-device SLMs, while preserving the domain-specific insights of each device. The experimental results show that Co-PLMs outperform state-of-the-art methods, achieving average increases of 5.38% in Rouge-L and 4.88% in EM.

cs.DC↗

arXiv2025

MetaSTH-Sleep: Towards Effective Few-Shot Sleep Stage Classification for Health Management with Spatial-Temporal Hypergraph Enhanced Meta-Learning

Accurate classification of sleep stages based on bio-signals is fundamental not only for automatic sleep stage annotation, but also for clinical health management and continuous sleep monitoring. Traditionally, this task relies on experienced clinicians to manually annotate data, a process that is both time-consuming and labor-intensive. In recent years, deep learning methods have shown promise in automating this task. However, three major challenges remain: (1) deep learning models typically require large-scale labeled datasets, making them less effective in real-world settings where annotated data is limited; (2) significant inter-individual variability in bio-signals often results in inconsistent model performance when applied to new subjects, limiting generalization; and (3) existing approaches often overlook the high-order relationships among bio-signals, failing to simultaneously capture signal heterogeneity and spatial-temporal dependencies. To address these issues, we propose MetaSTH-Sleep, a few-shot sleep stage classification framework based on spatial-temporal hypergraph enhanced meta-learning. Our approach enables rapid adaptation to new subjects using only a few labeled samples, while the hypergraph structure effectively models complex spatial interconnections and temporal dynamics simultaneously in EEG signals. Experimental results demonstrate that MetaSTH-Sleep achieves substantial performance improvements across diverse subjects, offering valuable insights to support clinicians in sleep stage annotation.

cs.LG↗

arXiv2025

Towards Heterogeneity-Aware and Energy-Efficient Topology Optimization for Decentralized Federated Learning in Edge Environment

Federated learning (FL) has emerged as a promising paradigm within edge computing (EC) systems, enabling numerous edge devices to collaboratively train artificial intelligence (AI) models while maintaining data privacy. To overcome the communication bottlenecks associated with centralized parameter servers, decentralized federated learning (DFL), which leverages peer-to-peer (P2P) communication, has been extensively explored in the research community. Although researchers design a variety of DFL approach to ensure model convergence, its iterative learning process inevitably incurs considerable cost along with the growth of model complexity and the number of participants. These costs are largely influenced by the dynamic changes of topology in each training round, particularly its sparsity and connectivity conditions. Furthermore, the inherent resources heterogeneity in the edge environments affects energy efficiency of learning process, while data heterogeneity degrades model performance. These factors pose significant challenges to the design of an effective DFL framework for EC systems. To this end, we propose Hat-DFed, a heterogeneity-aware and coset-effective decentralized federated learning (DFL) framework. In Hat-DFed, the topology construction is formulated as a dual optimization problem, which is then proven to be NP-hard, with the goal of maximizing model performance while minimizing cumulative energy consumption in complex edge environments. To solve this problem, we design a two-phase algorithm that dynamically constructs optimal communication topologies while unbiasedly estimating their impact on both model performance and energy cost. Additionally, the algorithm incorporates an importance-aware model aggregation mechanism to mitigate performance degradation caused by data heterogeneity.

cs.LG↗

arXiv2025

A Semi-Supervised Federated Learning Framework with Hierarchical Clustering Aggregation for Heterogeneous Satellite Networks

Low Earth Orbit (LEO) satellites are emerging as key components of 6G networks, with many already deployed to support large-scale Earth observation and sensing related tasks. Federated Learning (FL) presents a promising paradigm for enabling distributed intelligence in these resource-constrained and dynamic environments. However, achieving reliable convergence, while minimizing both processing time and energy consumption, remains a substantial challenge, particularly in heterogeneous and partially unlabeled satellite networks. To address this challenge, we propose a novel semi-supervised federated learning framework tailored for LEO satellite networks with hierarchical clustering aggregation. To further reduce communication overhead, we integrate sparsification and adaptive weight quantization techniques. In addition, we divide the FL clustering into two stages: satellite cluster aggregation stage and Ground Stations (GSs) aggregation stage. The supervised learning at GSs guides selected Parameter Server (PS) satellites, which in turn support fully unlabeled satellites during the federated training process. Extensive experiments conducted on a satellite network testbed demonstrate that our proposal can significantly reduce processing time (up to 3x) and energy consumption (up to 4x) compared to other comparative methods while maintaining model accuracy.

cs.DC↗

arXiv2025

Multimodal Fusion via Hypergraph Autoencoder and Contrastive Learning for Emotion Recognition in Conversation

Multimodal emotion recognition in conversation (MERC) seeks to identify the speakers' emotions expressed in each utterance, offering significant potential across diverse fields. The challenge of MERC lies in balancing speaker modeling and context modeling, encompassing both long-distance and short-distance contexts, as well as addressing the complexity of multimodal information fusion. Recent research adopts graph-based methods to model intricate conversational relationships effectively. Nevertheless, the majority of these methods utilize a fixed fully connected structure to link all utterances, relying on convolution to interpret complex context. This approach can inherently heighten the redundancy in contextual messages and excessive graph network smoothing, particularly in the context of long-distance conversations. To address this issue, we propose a framework that dynamically adjusts hypergraph connections by variational hypergraph autoencoder (VHGAE), and employs contrastive learning to mitigate uncertainty factors during the reconstruction process. Experimental results demonstrate the effectiveness of our proposal against the state-of-the-art methods on IEMOCAP and MELD datasets. We release the code to support the reproducibility of this work at https://github.com/yzjred/-HAUCL.

cs.MM↗

arXiv2025

DHLight: Multi-agent Policy-based Directed Hypergraph Learning for Traffic Signal Control

Recent advancements in Deep Reinforcement Learning (DRL) and Graph Neural Networks (GNNs) have demonstrated notable promise in the realm of intelligent traffic signal control, facilitating the coordination across multiple intersections. However, the traditional methods rely on standard graph structures often fail to capture the intricate higher-order spatio-temporal correlations inherent in real-world traffic dynamics. Standard graphs cannot fully represent the spatial relationships within road networks, which limits the effectiveness of graph-based approaches. In contrast, directed hypergraphs provide more accurate representation of spatial information to model complex directed relationships among multiple nodes. In this paper, we propose DHLight, a novel multi-agent policy-based framework that synergistically integrates directed hypergraph learning module. This framework introduces a novel dynamic directed hypergraph construction mechanism, which captures complex and evolving spatio-temporal relationships among intersections in road networks. By leveraging the directed hypergraph relational structure, DHLight empowers agents to achieve adaptive decision-making in traffic signal control. The effectiveness of DHLight is validated against state-of-the-art baselines through extensive experiments in various network datasets. We release the code to support the reproducibility of this work at https://github.com/LuckyVoasem/Traffic-Light-control

cs.MA↗

arXiv2025

Intelligent Task Management via Dynamic Multi-region Division in LEO Satellite Networks

As a key complement to terrestrial networks and a fundamental component of future 6G systems, Low Earth Orbit (LEO) satellite networks are expected to provide high-quality communication services when integrated with ground-based infrastructure, thereby attracting significant research interest. However, the limited satellite onboard resources and the uneven distribution of computational workloads often result in congestion along inter-satellite links (ISLs) that degrades task processing efficiency. Effectively managing the dynamic and large-scale topology of LEO networks to ensure balanced task distribution remains a critical challenge. To this end, we propose a dynamic multi-region division framework for intelligent task management in LEO satellite networks. This framework optimizes both intra- and inter-region routing to minimize task delay while balancing the utilization of computational and communication resources. Based on this framework, we propose a dynamic multi-region division algorithm based on the Genetic Algorithm (GA), which adaptively adjusts the size of each region based on the workload status of individual satellites. Additionally, we incorporate an adaptive routing algorithm and a task splitting and offloading scheme based on Multi-Agent Deep Deterministic Policy Gradient (MA-DDPG) to effectively accommodate the arriving tasks. Simulation results demonstrate that our proposed framework outperforms comparative methods in terms of the task delay, energy consumption per task, and task completion rate.

cs.DC↗

arXiv2025

CHASE: A Causal Hypergraph based Framework for Root Cause Analysis in Multimodal Microservice Systems

In recent years, the widespread adoption of distributed microservice architectures within the industry has significantly increased the demand for enhanced system availability and robustness. Due to the complex service invocation paths and dependencies in enterprise-level microservice systems, it is challenging to locate the anomalies promptly during service invocations, thus causing intractable issues for normal system operations and maintenance. In this paper, we propose a Causal Heterogeneous grAph baSed framEwork for root cause analysis, namely CHASE, for microservice systems with multimodal data, including traces, logs, and system monitoring metrics. Specifically, related information is encoded into representative embeddings and further modeled by a multimodal invocation graph. Following that, anomaly detection is performed on each instance node with attentive heterogeneous message passing from its adjacent metric and log nodes. Finally, CHASE learns from the constructed hypergraph with hyperedges representing the flow of causality and performs root cause localization. We evaluate the proposed framework on two public microservice datasets with distinct attributes and compare with the state-of-the-art methods. The results show that CHASE achieves the average performance gain up to 36.2%(A@1) and 29.4%(Percentage@1), respectively to its best counterpart.

cs.LG↗

arXiv2025

Towards Multi-agent Reinforcement Learning based Traffic Signal Control through Spatio-temporal Hypergraphs

Traffic signal control systems (TSCSs) are integral to intelligent traffic management, fostering efficient vehicle flow. Traditional approaches often simplify road networks into standard graphs, which results in a failure to consider the dynamic nature of traffic data at neighboring intersections, thereby neglecting higher-order interconnections necessary for real-time control. To address this, we propose a novel TSCS framework to realize intelligent traffic control. This framework collaborates with multiple neighboring edge computing servers to collect traffic information across the road network. To elevate the efficiency of traffic signal control, we have crafted a multi-agent soft actor-critic (MA-SAC) reinforcement learning algorithm. Within this algorithm, individual agents are deployed at each intersection with a mandate to optimize traffic flow across the road network collectively. Furthermore, we introduce hypergraph learning into the critic network of MA-SAC to enable the spatio-temporal interactions from multiple intersections in the road network. This method fuses hypergraph and spatio-temporal graph structures to encode traffic data and capture the complex spatio-temporal correlations between multiple intersections. Our empirical evaluation, tested on varied datasets, demonstrates the superiority of our framework in minimizing average vehicle travel times and sustaining high-throughput performance. This work facilitates the development of more intelligent urban traffic management solutions. We release the code to support the reproducibility of this work at https://github.com/Edun-Eyes/TSC

cs.MA↗

arXiv2025

CCRSat: A Collaborative Computation Reuse Framework for Satellite Edge Computing Networks

In satellite computing applications, such as remote sensing, tasks often involve similar or identical input data, leading to the same processing results. Computation reuse is an emerging paradigm that leverages the execution results of previous tasks to enhance the utilization of computational resources. While this paradigm has been extensively studied in terrestrial networks with abundant computing and caching resources, such as named data networking (NDN), it is essential to develop a framework appropriate for resource-constrained satellite networks, which are expected to have longer task completion times. In this paper, we propose CCRSat, a collaborative computation reuse framework for satellite edge computing networks. CCRSat initially implements local computation reuse on an independent satellite, utilizing a satellite reuse state (SRS) to assess the efficiency of computation reuse. Additionally, an inter-satellite computation reuse algorithm is introduced, which utilizes the collaborative sharing of similarity in previously processed data among multiple satellites. The evaluation results tested on real-world datasets demonstrate that, compared to comparative scenarios, our proposed CCRSat can significantly reduce task completion time by up to 62.1% and computational resource consumption by up to 28.8%.

cs.DC↗

arXiv2025

FedHC: A Hierarchical Clustered Federated Learning Framework for Satellite Networks

With the proliferation of data-driven services, the volume of data that needs to be processed by satellite networks has significantly increased. Federated learning (FL) is well-suited for big data processing in distributed, resource-constrained satellite environments. However, ensuring its convergence performance while minimizing processing time and energy consumption remains a challenge. To this end, we propose a hierarchical clustered federated learning framework, FedHC. This framework employs a satellite-clustered parameter server (PS) selection algorithm at the cluster aggregation stage, grouping nearby satellites into distinct clusters and designating a cluster center as the PS to accelerate model aggregation. Several communicable cluster PS satellites are then selected through ground stations to aggregate global parameters, facilitating the FL process. Moreover, a meta-learning-driven satellite re-clustering algorithm is introduced to enhance adaptability to dynamic satellite cluster changes. The extensive experiments on satellite networks testbed demonstrate that FedHC can significantly reduce processing time (up to 3x) and energy consumption (up to 2x) compared to other comparative methods while maintaining model accuracy.

cs.DC↗