Tao Chen

Consumer electronics devices face cybersecurity threats during data transmission. Federated Learning (FL) enables secure collaborative training by keeping data locally. However, FL faces data heterogeneity across diverse edge clients. Personalized FL (PFL) tackles this challenge, but existing approaches struggle to capture fine-grained features, limiting performance in real-world consumer electronics applications. This paper designs S-PFEL, a lightweight prototypical personalized federated edge learning framework for consumer mobile camera devices that balances security and personalization efficiency. S-PFEL introduces a dual-network architecture comprising an embedding network that extracts client-specific features and a hypernetwork that generates personalized models adapted to local data distributions on individual edge client devices. By integrating prototypical and contrastive learning, S-PFEL simultaneously captures class similarities and differences in the embedding space. Theoretical convergence guarantees are proved under the Lipschitz continuity and strong convexity assumptions. Experiments on SVHN, CIFAR-10/100, a facial emotion dataset, and a camera scene recognition dataset show S-PFEL outperforms six state-of-the-art baselines by up to 3.78% in non-IID scenarios, with robust performance across diverse deployment configurations.

The spread of multimodal misinformation demands verification that is both accurate and fast while keeping knowledge current. Large language models are powerful but costly and slow, and their static knowledge can lag behind events. We introduce EC-MFR, a hierarchical framework that divides work between edge and the cloud. The system first optionally decomposes the claim into a few targeted sub-claims to guide retrieval, retrieves text and image evidence, and then compresses it into a small set of question–answer items using a lightweight, quantized multimodal language model deployed at the edge. A compact verifier on the edge predicts a label with calibrated confidence. If confidence is high, the decision is returned immediately. If confidence is low, the claim is sent to the cloud where retrieval can be expanded and the reasoning can be redone by a stronger verifier. This design offers three core benefits. It makes reasoning explicit through question–answer items, which shortens prompts and improves auditability. It improves retrieval recall via a light decomposition step that produces targeted sub-queries. Finally, it lets most easy claims finish on the edge to reduce cost and latency while preserving accuracy on difficult claims by allowing the cloud to broaden evidence and refine reasoning. Experiments on MOCHEG and AVERITEC validate the approach. Notably, EC-MFR achieves highly competitive accuracy of 54.10% on the multimodal MOCHEG dataset, and reaches 68.80% on AVERITEC under realistic retrieval settings, outperforming the GPT-4o cloud-only baseline by 6.6 percentage points. Furthermore, system-level profiling on edge hardware demonstrates that EC-MFR reduces processing costs by 51.8% and accelerates inference latency by 2.4× for edge-resolved claims, confirming a highly favorable accuracy–efficiency trade-off compared to existing multimodal fact-checking systems. We also formalize routing and efficiency and analyze calibration and retrieval.

Large language models (LLMs) face difficulties in leveraging the syntactic structures and entity relations embedded in text for long-document information extraction. To address this issue, this paper proposes a generative extraction method integrating heterogeneous topology awareness and spatial alignment. The method first extracts syntactic and coreference information to construct a heterogeneous document graph and employs a mixture-of-experts network to decouple and encode multi-type topological features. A component orthogonal projection mechanism and a graph-text contrastive learning strategy are then utilized to align the extracted graph features to the underlying semantic space of the language model with high fidelity. Furthermore, Topology-Aware Encoder compresses the global features into fixed-length structural prompts to guide text generation. Experiments on the ACE2005, WikiEvents, and DuEE datasets demonstrated that the proposed method achieved state-of-the-art performance on information extraction tasks. Consequently, these results suggest that the proposed framework is a promising approach for complex information extraction across base LLMs of different scales.

Large language models (LLMs) have demonstrated significant capabilities in natural language processing (NLP), but they often encounter challenges in the medical domain. This can result in insufficient alignment between generated answers and user intent, as well as factual deviations. To address these issues, we propose Co-MedGraphRAG, a novel framework combining knowledge graph reasoning with large–small model collaboration, aimed at improving the structural grounding and interpretability of medical responses. The framework operates through a multi-stage collaborative mechanism to augment question answering. First, a large language model constructs a question-specific knowledge graph (KG) containing pending entities (denoted as “none”) to explicitly define known and unknown variables. Subsequently, a hybrid reasoning strategy is employed to populate the pending entities, thereby completing the question-specific knowledge graph. Finally, this graph serves as critical structured evidence, combined with the original question, to augment the large language model in generating the final answer, implemented using Qwen2.5-7B and GLM4-9B in this paper. To evaluate the generated answers, we introduce a larger-parameter LLM(GPT-4o) to assess performance across five dimensions and compute an overall score. Experiments on three medical datasets demonstrate that Co-MedGraphRAG achieves consistent improvements in relevance, practicality, and structured knowledge support compared with mainstream Retrieval-Augmented Generation (RAG) frameworks. This work serves as a reference for researchers and developers designing medical question-answering frameworks and exploring decision-support applications.

Social media is flooded with disinformation about outbreaks, and emotional and ideological appeals can significantly influence public opinion more than objective facts. This study delves into the cognitive and affective responses of the public when confronted with disinformation about the COVID-19 epidemic on Weibo. To cope with the complexity of emotions, we designed a multidimensional complex affective space to depict the intricate cognition of the public. Drawing from human thinking processes, we proposed a fine-tuning method based on the large language model ChatGLM2-6B to augment semantic information and bolster the sentiment analysis capabilities of the large language model. The experimental results indicated that incorporating semantic information enhances the prediction accuracy of the large language model, with varying effects across different sentiment dimensions. Moreover, we analyze the public’s topics concerning epidemic disinformation, and the findings demonstrate that the public tends to generate similar topics for such fake news. Emojis have emerged as a means for individuals to supplement the expression of complex emotions.