检索结果-内蒙古大学图书馆

Privacy-preserving recommendation with coarse-grained spatiotemporal contexts

science China(information sciences) 2025年第4期68卷 66-81页

作者： Lei CHEN Chen GAO Jiahuan LEI Xiaoyi DU Xinlei SHI Hengliang LUO Depeng JIN Yong LI Meng WANG Department of Electronic Engineering BNRist Tsinghua University Meituan Inc. School of Computer Science and Information Engineering Hefei University of Technology

The behavior of users on online life service platforms like Meituan and Yelp often occurs within specific finegrained spatiotemporal contexts(i.e., when and where). Recommender systems, designed to serve millions of users, typically operate in a fully server-based manner, requiring on-device users to upload their behavioral data, including fine-grained spatiotemporal contexts, to the server, which has sparked public concern regarding privacy. Consequently, user devices only upload coarse-grained spatiotemporal contexts for user privacy protection. However, previous research mostly focuses on modeling fine-grained spatiotemporal contexts using knowledge graph convolutional models, which are not applicable to coarse-grained spatiotemporal contexts in privacy-constrained recommender systems. In this paper, we investigate privacy-preserving recommendation by leveraging coarse-grained spatiotemporal contexts. We propose the coarse-grained spatiotemporal knowledge graph for privacy-preserving recommendation(CSKG), which explicitly models spatiotemporal co-occurrences using common-sense knowledge from coarse-grained contexts. Specifically, we begin by constructing a spatiotemporal knowledge graph tailored to coarse-grained spatiotemporal contexts. Then we employ a learnable metagraph network that integrates common-sense information to filter and extract co-occurrences. CSKG evaluates the impact of coarsegrained spatiotemporal contexts on user behavior through the use of a knowledge graph convolutional network. Finally, we introduce joint learning to effectively learn representations. By conducting experiments on two real large-scale datasets,we achieve an average improvement of about 11.0% on two ranking metrics. The results clearly demonstrate that CSKG outperforms state-of-the-art baselines.

关键词： privacy-preserveing coarse-grained spatiotemporal contexts recommender systems

来源：评论

学校读者我要写书评

暂无评论

A recover-then-discriminate framework for robust anomaly detection

引用

science China(information sciences) 2025年第4期68卷 300-318页

作者： Peng XING Dong ZHANG Jinhui TANG Zechao LI School of Computer Science and Engineering Nanjing University of Science and Technology Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology

Anomaly detection(AD) has been extensively studied and applied across various scenarios in recent years. However, gaps remain between the current performance and the desired recognition accuracy required for practical *** paper analyzes two fundamental failure cases in the baseline AD model and identifies key reasons that limit the recognition accuracy of existing approaches. Specifically, by Case-1, we found that the main reason detrimental to current AD methods is that the inputs to the recovery model contain a large number of detailed features to be recovered, which leads to the normal/abnormal area has not/has been recovered into its original state. By Case-2, we surprisingly found that the abnormal area that cannot be recognized in image-level representations can be easily recognized in the feature-level representation. Based on the above observations, we propose a novel recover-then-discriminate(ReDi) framework for *** takes a self-generated feature map(e.g., histogram of oriented gradients) and a selected prompted image as explicit input information to address the identified in Case-1. Additionally, a feature-level discriminative network is introduced to amplify abnormal differences between the recovered and input representations. Extensive experiments on two widely used yet challenging AD datasets demonstrate that ReDi achieves state-of-the-art recognition accuracy.

关键词： recovery network HOG prompt discriminative network self-correlation loss anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Enhanced Acceleration for Generalized Nonconvex Low-Rank Matrix Learning

引用

Chinese Journal of electronics 2025年第1期34卷 98-113页

作者： Hengmin Zhang Jian Yang Wenli Du Bob Zhang Zhiyuan Zha Bihan Wen School of Electrical and Electronic Engineering Nanyang Technological University School of Computer Science and Engineering Nanjing University of Science and Technology School of Information Science and Engineering East China University of Science and Technology Department of Electrical and Computer Engineering University of Macau

Matrix minimization techniques that employ the nuclear norm have gained recognition for their applicability in tasks like image inpainting, clustering, classification, and reconstruction. However, they come with inherent biases and computational burdens, especially when used to relax the rank function, making them less effective and efficient in real-world scenarios. To address these challenges, our research focuses on generalized nonconvex rank regularization problems in robust matrix completion, low-rank representation, and robust matrix regression. We introduce innovative approaches for effective and efficient low-rank matrix learning, grounded in generalized nonconvex rank relaxations inspired by various substitutes for the ?0-norm relaxed functions. These relaxations allow us to more accurately capture low-rank structures. Our optimization strategy employs a nonconvex and multi-variable alternating direction method of multipliers, backed by rigorous theoretical analysis for complexity and *** algorithm iteratively updates blocks of variables, ensuring efficient convergence. Additionally, we incorporate the randomized singular value decomposition technique and/or other acceleration strategies to enhance the computational efficiency of our approach, particularly for large-scale constrained minimization problems. In conclusion, our experimental results across a variety of image vision-related application tasks unequivocally demonstrate the superiority of our proposed methodologies in terms of both efficacy and efficiency when compared to most other related learning methods.

关键词： Learning systems Image recognition Minimization Computational efficiency Complexity theory Matrix decomposition Optimization Image reconstruction Singular value decomposition Convergence

来源：评论

学校读者我要写书评

暂无评论

MMInstruct: a high-quality multi-modal instruction tuning dataset with extensive diversity

引用

science China(information sciences) 2024年第12期67卷 36-51页

作者： Yangzhou LIU Yue CAO Zhangwei GAO Weiyun WANG Zhe CHEN Wenhai WANG Hao TIAN Lewei LU Xizhou ZHU Tong LU Yu QIAO Jifeng DAI School of Computer Science Nanjing University School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai AI Laboratory School of Computer Science Fudan University Department of Information Engineering The Chinese University of Hong Kong SenseTime Research Department of Electronic Engineering Tsinghua University

Despite the effectiveness of vision-language supervised fine-tuning in enhancing the performance of vision large language models(VLLMs), existing visual instruction tuning datasets include the following limitations.(1) Instruction annotation quality: despite existing VLLMs exhibiting strong performance,instructions generated by those advanced VLLMs may still suffer from inaccuracies, such as hallucinations.(2) Instructions and image diversity: the limited range of instruction types and the lack of diversity in image data may impact the model's ability to generate diversified and closer to real-world scenarios outputs. To address these challenges, we construct a high-quality, diverse visual instruction tuning dataset MMInstruct,which consists of 973k instructions from 24 domains. There are four instruction types: judgment, multiplechoice, long visual question answering, and short visual question answering. To construct MMInstruct, we propose an instruction generation data engine that leverages GPT-4V, GPT-3.5, and manual correction. Our instruction generation engine enables semi-automatic, low-cost, and multi-domain instruction generation at 1/6 the cost of manual construction. Through extensive experiment validation and ablation experiments,we demonstrate that MMInstruct could significantly improve the performance of VLLMs, e.g., the model fine-tuning on MMInstruct achieves new state-of-the-art performance on 10 out of 12 benchmarks. The code and data shall be available at https://***/yuecao0119/MMInstruct.

关键词： instruction tuning multi-modal multi-domain dataset vision large language model

来源：评论

学校读者我要写书评

暂无评论

DocPedia: unleashing the power of large multimodal model in the frequency domain for versatile document understanding

引用

science China(information sciences) 2024年第12期67卷 65-78页

作者： Hao FENG Qi LIU Hao LIU Jingqun TANG Wengang ZHOU Houqiang LI Can HUANG Department of Electronic Engineering and Information Science University of Science and Technology of China ByteDance Inc.

In this work, we present DocPedia, a novel large multimodal model(LMM) for versatile OCRfree document understanding, capable of parsing images up to 2560 × 2560 resolution. Unlike existing studies that either struggle with high-resolution documents or give up the large language model thus vision or language ability constrained, our DocPedia directly processes visual input in the frequency domain rather than the pixel space. The unique characteristic enables DocPedia to capture a greater amount of visual and textual information using a limited number of visual tokens. To consistently enhance both the perception and comprehension abilities of our DocPedia, we develop a dual-stage training strategy and enrich instructions/annotations of all training tasks covering multiple document types. Extensive quantitative and qualitative experiments are conducted on various publicly available benchmarks and the results confirm the mutual benefits of jointly learning perception and comprehension tasks. The results provide further evidence of the effectiveness and superior performance of our DocPedia over other methods.

关键词： document understanding large multimodal model OCR-free high-resolution frequency

来源：评论

学校读者我要写书评

暂无评论

Data delivery delay and cross-layer packet size analysis for reliable transmission of Licklider transmission protocol in space networks

引用

science China(information sciences) 2024年第9期67卷 229-244页

作者： Guannan YANG Ruhai WANG Kanglian ZHAO Wenfeng LI Dong YAN School of Information Engineering Nanjing University of Finance and Economics Drayer Department of Electrical and Computer Engineering Lamar University School of Electronic Science and Engineering Nanjing University Beijing Aircraft Overall Design Department

Delay/disruption tolerant networking(DTN) is proposed as a networking architecture to overcome challenging space communication characteristics for reliable data transmission service in presence of long propagation delays and/or lengthy link disruptions. Bundle protocol(BP) and Licklider Transmission Protocol(LTP) are the main key technologies for DTN. LTP red transmission offers a reliable transmission mechanism for space networks. One of the key metrics used to measure the performance of LTP in space applications is the end-to-end data delivery delay, which is influenced by factors such as the quality of spatial channels and the size of cross-layer packets. In this paper, an end-to-end reliable data delivery delay model of LTP red transmission is proposed using a roulette wheel algorithm, and the roulette wheel algorithm is more in line with the typical random characteristics in space networks. The proposed models are validated through real data transmission experiments on a semi-physical testing platform. Furthermore, the impact of cross-layer packet size on the performance of LTP reliable transmission is analyzed, with a focus on bundle size, block size, and segment size. The analysis and study results presented in this paper offer valuable contributions towards enhancing the reliability of LTP transmission in space communication scenarios.

关键词： Licklider transmission protocol DTN bundle protocol cross-layer packets size space communication

来源：评论

学校读者我要写书评

暂无评论

Self-attention reinforcement learning for multi-beam combining in mmW ave 3D-MIMO systems

引用

science China(information sciences) 2023年第6期66卷 204-221页

作者： Yingzhi HUANG Zhaoyang ZHANG Jingze CHE Zhaohui YANG Qianqian YANG Kai-Kit WONG College of Information Science and Electronic Engineering Zhejiang University Department of Electronic and Electrical Engineering University College London

Machine learning(ML)has been empowering all aspects of the wireless communication system design, among which, the reinforcement learning(RL)-based approaches have attracted a lot of research attention since they can interact with the environment directly and learn from the collected experiences efficiently. In this paper, we propose a novel and efficient RL-based multi-beam combining scheme for future millimeter-wave(mmWave)three-dimensional(3D)multi-input multi-output(MIMO)communication systems. The proposed scheme does not require perfect channel state information(CSI)or precise user location information which both are generally difficult to obtain in practice, and well addresses the crucial challenge of computational complexity incurred by the extremely huge state and action spaces associated with multiple users, multiple paths, and multiple 3D beams. In particular, a self-attention deep deterministic policy gradient(DDPG)-based beam selection and combination framework is proposed to learn the 3D beamforming pattern without CSI adaptively. We aim to maximize the sum-rate of the mmWave 3D-MIMO system by optimizing the serving beam set and the corresponding combining weights for each user. To this end, the transformer is incorporated into the DDPG to obtain the global information of the input elements and capture the signal directions precisely, which leads to a near-optimal beamformer design. Simulation results verify the superiority of the proposed self-attention DDPG over conventional ML-based beamforming schemes in terms of sum-rate under various scenarios.

关键词： reinforcement learning (RL) deep deterministic policy gradient (DDPG) self-attention precoding/combining millimeter-wave (mmWave) multi-input multi-output (MIMO)

来源：评论

学校读者我要写书评

暂无评论

Open-Space Emergency Guiding with Individual Density Prediction Based on Internet of Things Localization

引用

IEEE Transactions on Emerging Topics in Computational Intelligence 2025年第1期9卷 785-797页

作者： Chen, Lien-Wu Huang, Hao-Wei Chen, Yi-Ju Tsai, Ming-Fong Feng Chia University Department of Information Engineering and Computer Science Taichung407 Taiwan National United University Department of Electronic Engineering Miaoli360 Taiwan

This article proposes an open-space emergency guiding (OSEG) framework that explores deep learning techniques to predict individual densities for evacuation based on Internet of Things localization. The OSEG framework adopts Densely Connected Convolutional Networks that can reserve fine-grained features in earlier convolutional layers and extract them in later convolutional layers to efficiently reduce prediction errors. In addition, OSEG can dynamically guide individuals to exits in the shortest time though balancing the evacuation load among exits as individual densities are non-uniform in indoor spaces. According to our review of relevant research, this is the first framework that integrates deep learning for individual density prediction with load-balancing evacuation for emergency guiding in open indoor spaces, where the distance to each exit, capacities of all exits, concurrent moving of evacuees, and distribution of indoor individuals/groups are considered simultaneously. Simulation results show that our approach outperforms existing methods and can balance the evacuation load of each exit for individuals/groups to minimize total evacuation time, which can reduce 50% of total evacuation time at most compared with existing methods. © 2017 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Modern Machine Learning Solution for Electricity Consumption Management in Smart Buildings

引用

IEEE engineering Management Review 2025年第1期53卷 54-62页

作者： Gautam, Sandeep Kumar Shrivastava, Vinayak Udmale, Sandeep S. Singh, Amit Kumar Singh, Sanjay Kumar Department of Computer Science and Engineering Varanasi221005 India Department of Computer Engineering and Information Technology Mumbai400019 India National Institute of Technology Department of Computer Science and Engineering Patna800005 India

Effective management of electricity consumption (EC) in smart buildings (SBs) is crucial for optimizing operational efficiency, cost savings, and ensuring sustainable resource utilization. Accurate EC prediction enables proactive decision-making, ensuring that resources are allocated efficiently to meet actual demand levels while maintaining occupant comfort. Population growth, building expansion, and technology usage swiftly escalate electricity demand, thus necessitating economical EC management strategies and assist consumers to better understand and strategically plan their EC. To address these challenges, this article proposes a novel approach based on a hybrid prediction model combining temporal convolutional networks (TCNs) and gated recurrent units (GRU). This approach capitalizes on the strengths of both TCN and GRU. TCN is adept at efficiently identifying diverse patterns, particularly within the complex working environments of SBs, by effectively capturing high- and low-frequency information. Subsequently, GRU is leveraged to address the long-term dependencies within the data, enhancing the accuracy of EC prediction. In this article, the results demonstrate the effectiveness of the proposed hybrid model, outperforming competitive methods with an impressive mean absolute error score. This underscores the potential of this approach to improve energy management practices significantly within SB environments, ultimately enhancing both operational efficiency and occupant satisfaction. © 1973-2011 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Subchannel selection methods for 3GPP C-V2X networks by considering vehicular mobility

引用

Telecommunication Systems 2024年第3期86卷 503-518页

作者： Pan, Meng-Shiuan Kao, Shao-Wei Department of Electronic Engineering National Taipei University of Technology Taipei Taiwan Department of Computer Science and Information Engineering Tamkang University New Taipei Taiwan

The 3GPP vehicle-to-everything (C-V2X) technology is a key solution to provide communication services for applications of intelligent transportation systems (ITS). According to the C-V2X specification, vehicles are allowed to communicate with each other without the help of cellular networks and can select subchannels (i.e. radio resources) by themselves. However, this design will induce subchannel collisions because of the mobility of vehicles. This work aims to relieve the aforementioned problem. We propose a cooperative subchannel selection (CoSS) scheme and a localized subchannel selection (LoSS) scheme, which consider vehicular mobility to prevent subchannel collisions. Based on the collected information, CoSS predicts possible subchannel collisions due to vehicles’ location changes. On the other hand, LoSS avoids subchannel collisions by organizing the usage of subchannels based on vehicles’ relative moving speeds. The simulation results indicate that the proposed schemes can effectively reduce subchannel collision ratio and increase packet delivery ratio for the network. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Vehicle to Everything

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：