检索结果-内蒙古大学图书馆

Bridging Trajectory-Aware Evolutionary Graph Learning and Large Language Models for Enhancing Navigability in Social Internet of Things

学校读者我要写书评

暂无评论

IEEE Internet of Things Journal 2025年

作者： Bi, Xin Han, Zhubin Yao, Xin Zhao, Xiangguo Wang, Yu-Ping Yuan, Ye State Key Laboratory of Intelligent Deep Metal Mining and Equipment China Northeastern University School of Computer Science and Engineering Shenyang110819 China Northeastern University College of Software Shenyang110819 China Beijing Institute of Technology School of Computer Science and Technology Beijing100091 China

Social Internet of Things (SIoT) has emerged as a novel paradigm that enhances IoT service capabilities by leveraging device-level social relationships. However, the explosive growth of heterogeneous devices, the dynamic mobile device trajectories, and complex spatiotemporal interaction patterns severely hinder SIoT network navigability. Particularly, the device mobility and contextual diversity pose significant challenges to social relation classification, a critical task for efficient routing and service discovery. Existing approaches, primarily designed for static or homogeneous networks, fail to adequately capture the evolving contextual dependencies and the spatiotemporal heterogeneity in SIoT. To address these challenges, we propose Trajectory-Aware Graph LLM (TAGLLM), a novel framework that enhances SIoT navigability through context-aware relation classification. TAGLLM introduces a multi-feature fusion trajectory evolutionary graph encoder to jointly model complex device attributes, social relations, and dynamic trajectories. Furthermore, a structural graph-text token alignment strategy is designed to exploit the generalization ability and contextual understanding capabilities of Large Language Models (LLMs), enabling more effective modeling of heterogeneous and dynamic SIoT scenarios. Extensive experiments on real-world SIoT datasets demonstrate that TAGLLM outperforms state-of-the-art baselines across multiple evaluation metrics, highlighting its potential to push the frontier of graph learning and LLM integration in SIoT applications. © 2014 IEEE.

关键词： Network theory (graphs)

When Neural Code Completion Models Size up the Situation: Attaining Cheaper and Faster Completion through Dynamic Model Inference

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Sun, Zhensu Du, Xiaoning Song, Fu Wang, Shangwen Li, Li Monash University MelbourneVIC Australia State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China National University of Defense Technology Changsha China Beihang University Yunnan Key Laboratory of Software Engineering Beijing China

Leveraging recent advancements in large language models, modern neural code completion models have demonstrated the capability to generate highly accurate code suggestions. However, their massive size poses challenges in terms of computational costs and environmental impact, hindering their widespread adoption in practical scenarios. Dynamic inference emerges as a promising solution, as it allocates minimal computation during inference while maintaining the model’s performance. In this research, we explore dynamic inference within the context of code completion. Initially, we conducted an empirical investigation on GPT-2, focusing on the inference capabilities of intermediate layers for code completion. We found that 54.4% of tokens can be accurately generated using just the first layer, signifying significant computational savings potential. Moreover, despite using all layers, the model still fails to predict 14.5% of tokens correctly, and the subsequent completions continued from them are rarely considered helpful, with only a 4.2% Acceptance Rate. These findings motivate our exploration of dynamic inference in code completion and inspire us to enhance it with a decision-making mechanism that stops the generation of incorrect code. We thus propose a novel dynamic inference method specifically tailored for code completion models. This method aims not only to produce correct predictions with largely reduced computation but also to prevent incorrect predictions proactively. Our extensive evaluation shows that it can averagely skip 1.7 layers out of 16 layers in the models, leading to an 11.2% speedup with only a marginal 1.1% reduction in ROUGE-L. Copyright © 2024, The Authors. All rights reserved.

关键词： Forecasting

ODRL: A Benchmark for Off-Dynamics Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lyu, Jiafei Xu, Kang Xu, Jiacheng Yan, Mengbei Yang, Jingwen Zhang, Zongzhang Bai, Chenjia Lu, Zongqing Li, Xiu Tsinghua Shenzhen International Graduate School Tsinghua University China Tencent China National Key Laboratory for Novel Software Technology Nanjing University China China Telecom China School of Computer Science Peking University China

We consider off-dynamics reinforcement learning (RL) where one needs to transfer policies across different domains with dynamics mismatch. Despite the focus on developing dynamics-aware algorithms, this field is hindered due to the lack of a standard benchmark. To bridge this gap, we introduce ODRL, the first benchmark tailored for evaluating off-dynamics RL methods. ODRL contains four experimental settings where the source and target domains can be either online or offline, and provides diverse tasks and a broad spectrum of dynamics shifts, making it a reliable platform to comprehensively evaluate the agent’s adaptation ability to the target domain. Furthermore, ODRL includes recent off-dynamics RL algorithms in a unified framework and introduces some extra baselines for different settings, all implemented in a single-file manner. To unpack the true adaptation capability of existing methods, we conduct extensive benchmarking experiments, which show that no method has universal advantages across varied dynamics shifts. We hope this benchmark can serve as a cornerstone for future research endeavors. Our code is publicly available at https://***/OffDynamicsRL/off-dynamics-rl. © 2024, CC BY.

关键词： Reinforcement learning

Newton-cotes graph neural networks: on the time evolution of dynamic systems 23

学校读者我要写书评

暂无评论

Newton-cotes graph neural networks: on the time evolution of...

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Lingbing Guo Weiqing Wang Zhuo Chen Ningyu Zhang Zequn Sun Yixuan Lai Qiang Zhang Huajun Chen College of Computer Science and Technology Zhejiang University and Zhejiang University - Ant Group Joint Laboratory of Knowledge Graph and ZJU-Hangzhou Global Scientific and Technological Innovation Center Department of Data Science & AI Monash University College of Computer Science and Technology Zhejiang University and Zhejiang University - Ant Group Joint Laboratory of Knowledge Graph State Key Laboratory for Novel Software Technology Nanjing University College of Computer Science and Technology Zhejiang University and ZJU-Hangzhou Global Scientific and Technological Innovation Center

Reasoning system dynamics is one of the most important analytical approaches for many scientific studies. With the initial state of a system as input, the recent graph neural networks (GNNs)-based methods are capable of predicting the future state distant in time with high accuracy. Although these methods have diverse designs in modeling the coordinates and interacting forces of the system, we show that they actually share a common paradigm that learns the integration of the velocity over the interval between the initial and terminal coordinates. However, their integrand is constant w.r.t. time. Inspired by this observation, we propose a new approach to predict the integration based on several velocity estimations with Newton-Cotes formulas and prove its effectiveness theoretically. Extensive experiments on several benchmarks empirically demonstrate consistent and significant improvement compared with the state-of-the-art methods.

关键词：

A Privacy Preserving Method for IoT Forensics

学校读者我要写书评

暂无评论

A Privacy Preserving Method for IoT Forensics

IEEE Conference on Global Communications (GLOBECOM)

作者： Wenzheng Zhang Boxi Chen Xiao Fu Qing Gu Jin Shi Xiaojiang Du Xiaoyang Zhou State Key Laboratory for Novel Software Technology Nanjing University Nanjing China School of Information Management Nanjing University Nanjing China Department of Electrical and Computer Engineering Stevens Institute of Technology Hoboken NJ USA China Mobile Zijin(Jiangsu)Innovation Research Institute Nanjing China

ISBN: (数字)9798350351255

ISBN: (纸本)9798350351262

The diversity of the Internet of Things (IoT) poses challenges to privacy protection, especially in the field of digital forensics. How to ensure that only the private information of the suspect is provided, and not the irrelevant information of other users is disclosed is crucial, especially when obtaining evidence in the complex IoT environment. To the best of our knowledge, there are few studies on protecting the privacy of irrelevant users in the IoT forensics. However, it is very important to ensure that the evidence does not violate the privacy of other users when collecting evidence, because it directly determines whether the evidence is legal and whether it can be admissible in court. In this paper, a new method based on data provenance graph is designed to solve the privacy protection problem of IoT forensics. The key idea of this method is to protect privacy by dividing multi-user information and protecting it from an encryption perspective. The method consists of three main phrases: distinguishing different users' data provenance graphs using traversal search, node abstraction, and hiding techniques, utilizing pseudo-random dual-key negotiation methods tailored for the scenario to enhance privacy protection for unrelated users, and employing identity authentication technology to facilitate better investigation and extraction of data provenance graph information of criminal accomplices in specific scenarios. Example proves that this method has practical significance and promising application prospects in protecting the privacy of unrelated users in IoT forensics while ensuring evidence accessibility in special criminal scenarios.

关键词： Privacy Data privacy Digital forensics Semantics Authentication Smart homes Encryption Internet of Things Protection Monitoring

Orthogonal Annotation Benefits Barely-supervised Medical Image Segmentation

学校读者我要写书评

暂无评论

Orthogonal Annotation Benefits Barely-supervised Medical Ima...

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Heng Cai Shumeng Li Lei Qi Qian Yu Yinghuan Shi Yang Gao Nanjing University State Key Laboratory for Novel Software Technology and National Institute of Healthcare Data Science Nanjing University China Southest University Shandong Women's University

Recent trends in semi-supervised learning have significantly boosted the performance of 3D semi-supervised medical image segmentation. Compared with 2D images, 3D medical volumes involve information from different directions, e.g., transverse, sagittal, and coronal planes, so as to naturally provide complementary views. These complementary views and the intrinsic similarity among adjacent 3D slices inspire us to develop a novel annotation way and its corresponding semi-supervised model for effective segmentation. Specifically, we firstly propose the orthogonal annotation by only labeling two orthogonal slices in a labeled volume, which significantly relieves the burden of annotation. Then, we perform registration to obtain the initial pseudo labels for sparsely labeled volumes. Subsequently, by introducing unlabeled volumes, we propose a dual-network paradigm named Dense-Sparse Co-training (DeSCO) that exploits dense pseudo labels in early stage and sparse labels in later stage and meanwhile forces consistent output of two networks. Experimental results on three benchmark datasets validated our effectiveness in performance and efficiency in annotation. For example, with only 10 annotated slices, our method reaches a Dice up to 86.93% on KiTS19 dataset. Our code and models are available at https://***/HengCai-NJU/DeSCO.

关键词：

Unleashing the Power of Large Language Model for Denoising Recommendation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Shuyao Zheng, Zhi Sui, Yongduo Xiong, Hui School of Data Science University of Science and Technology of China Hefei China State Key Laboratory of Cognitive Intelligence University of Science and Technology of China Hefei China Tencent Shenzhen China Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong

Recommender systems are crucial for personalizing user experiences but often depend on implicit feedback data, which can be noisy and misleading. Existing denoising studies involve incorporating auxiliary information or learning strategies from interaction data. However, they struggle with the inherent limitations of external knowledge and interaction data, as well as the non-universality of certain predefined assumptions, hindering accurate noise identification. Recently, large language models (LLMs) have gained attention for their extensive world knowledge and reasoning abilities, yet their potential in enhancing denoising in recommendations remains underexplored. In this paper, we introduce LLaRD, a framework leveraging LLMs to improve denoising in recommender systems, thereby boosting overall recommendation performance. Specifically, LLaRD generates denoising-related knowledge by first enriching semantic insights from observational data via LLMs and inferring user-item preference knowledge. It then employs a novel Chain-of-Thought (CoT) technique over user-item interaction graphs to reveal relation knowledge for denoising. Finally, it applies the Information Bottleneck (IB) principle to align LLM-generated denoising knowledge with recommendation targets, filtering out noise and irrelevant LLM knowledge. Empirical results demonstrate LLaRD’s effectiveness in enhancing denoising and recommendation accuracy. The code is available at https://***/shuyao-wang/LLaRD. Copyright © 2025, The Authors. All rights reserved.

关键词：

Newton-Cotes Graph Neural Networks: On the Time Evolution of Dynamic Systems

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Guo, Lingbing Wang, Weiqing Chen, Zhuo Zhang, Ningyu Sun, Zequn Lai, Yixuan Zhang, Qiang Chen, Huajun College of Computer Science and Technology Zhejiang University China Zhejiang University Ant Group Joint Laboratory of Knowledge Graph China ZJU-Hangzhou Global Scientific and Technological Innovation Center China Department of Data Science & AI Monash University Australia State Key Laboratory for Novel Software Technology Nanjing University China

关键词： Graph neural networks

PVBF: A Framework for Mitigating Parameter Variation Imbalance in Online Continual Learning

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Tao, Zelin Deng, Hao Liu, Mingqing Zhang, Lijun Zhao, Shengjie School of Computer Science and Technology Tongji University Shanghai201804 China College of Electronic and Information Engineering Tongji University Shanghai201804 China National Key Laboratory for Novel Software Technology Nanjing University Nanjing210023 China Engineering Research Center of Key Software Technologies for Smart City Perception and Planning Ministry of Education Shanghai201804 China

Online continual learning (OCL), which enables AI systems to adaptively learn from non-stationary data streams, is commonly achieved using experience replay (ER)-based methods that retain knowledge by replaying stored past during training. However, these methods face challenges of prediction bias, stemming from deviations in parameter update directions during task transitions. This paper identifies parameter variation imbalance as a critical factor contributing to prediction bias in ER-based OCL. Specifically, using the proposed parameter variation evaluation method, we highlight two types of imbalance: correlation-induced imbalance, where certain parameters are disproportionately updated across tasks, and layer-wise imbalance, where output layer parameters update faster than those in preceding layers. To mitigate the above imbalances, we propose the Parameter Variation Balancing Framework (PVBF), which incorporates: 1) a novel method to compute parameter correlations with previous tasks based on parameter variations, 2) an encourage-and-consolidate (E&C) method utilizing parameter correlations to perform gradient adjustments across all parameters during training, 3) a dual-layer copy weights with reinit (D-CWR) strategy to slowly update output layer parameters for frequently occuring sample categories. Experiments on short and long task sequences demonstrate that PVBF significantly reduces prediction bias and improves OCL performance, achieving up to 47% higher accuracy compared to existing ER-based methods. Copyright © 2025, The Authors. All rights reserved.

关键词： Contrastive Learning