检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Morita, Shogo Zhang, Yan Yamauchi, Takuto Chen, Sinan Li, Jialong Tei, Kenji School of Computing Tokyo Institute of Technology Tokyo Japan International Research Center for Neurointelligence the University of Tokyo Tokyo Japan Department of Computer Science and Engineering Waseda University Tokyo Japan Center of Mathematical and Data Sciences Kobe University Kobe Japan

People with color vision deficiency often face challenges in distinguishing colors such as red and green, which can complicate daily tasks and require the use of assistive tools or environmental adjustments. Current support tools mainly focus on presentation-based aids, like the color vision modes found in iPhone accessibility settings. However, offering context-aware support, like indicating the doneness of meat, remains a challenge since task-specific solutions are not cost-effective for all possible scenarios. To address this, our paper proposes an application that provides contextual and autonomous assistance. This application is mainly composed of: (i) an augmented reality interface that efficiently captures context;and (ii) a multi-modal large language model-based reasoner that serves to cognitize the context and then reason about the appropriate support contents. Preliminary user experiments with two color vision deficient users across five different scenarios have demonstrated the effectiveness and universality of our application. © 2024, CC BY.

关键词： Color

来源：评论

学校读者我要写书评

暂无评论

Securely Fine-tuning Pre-trained Encoders Against Adversarial Examples

arXiv

引用

arXiv 2024年

作者： Zhou, Ziqi Li, Minghui Liu, Wei Hu, Shengshan Zhang, Yechao Wan, Wei Xue, Lulu Zhang, Leo Yu Yao, Dezhong Jin, Hai National Engineering Research Center for Big Data Technology and System China Services Computing Technology and System Lab. China Cluster and Grid Computing Lab. China Hubei Engineering Research Center on Big Data Security China Hubei Key Laboratory of Distributed System Security China School of Computer Science and Technology Huazhong University of Science and Technology China School of Software Engineering Huazhong University of Science and Technology China School of Cyber Science and Engineering Huazhong University of Science and Technology China School of Information and Communication Technology Griffith University Australia

With the evolution of self-supervised learning, the pre-training paradigm has emerged as a predominant solution within the deep learning landscape. Model providers furnish pre-trained encoders designed to function as versatile feature extractors, enabling downstream users to harness the benefits of expansive models with minimal effort through fine-tuning. Nevertheless, recent works have exposed a vulnerability in pre-trained encoders, highlighting their susceptibility to downstream-agnostic adversarial examples (DAEs) meticulously crafted by attackers. The lingering question pertains to the feasibility of fortifying the robustness of downstream models against DAEs, particularly in scenarios where the pre-trained encoders are publicly accessible to the attackers. In this paper, we initially delve into existing defensive mechanisms against adversarial examples within the pre-training paradigm. Our findings reveal that the failure of current defenses stems from the domain shift between pre-training data and downstream tasks, as well as the sensitivity of encoder parameters. In response to these challenges, we propose Genetic Evolution-Nurtured Adversarial Fine-tuning (Gen-AF), a two-stage adversarial fine-tuning approach aimed at enhancing the robustness of downstream models. Gen-AF employs a genetic-directed dual-track adversarial fine-tuning strategy in its first stage to effectively inherit the pre-trained encoder. This involves optimizing the pre-trained encoder and classifier separately while incorporating genetic regularization to preserve the model’s topology. In the second stage, Gen-AF assesses the robust sensitivity of each layer and creates a dictionary, based on which the top-k robust redundant layers are selected with the remaining layers held fixed. Upon this foundation, we conduct evolutionary adaptability fine-tuning to further enhance the model’s generalizability. Our extensive experiments, conducted across ten self-supervised training methods and six

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

A Survey of Attacks on Large Vision-Language Models: Resources, Advances, and Future Trends

arXiv

引用

arXiv 2024年

作者： Liu, Daizong Yang, Mingyu Qu, Xiaoye Zhou, Pan Cheng, Yu Hu, Wei Wangxuan Institute of Computer Technology Peking University No. 128 Zhongguancun North Street Beijing China The Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan430074 China The Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong

With the significant development of large models in recent years, Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across a wide range of multimodal understanding and reasoning tasks. Compared to traditional Large Language Models (LLMs), LVLMs present great potential and challenges due to its closer proximity to the multi-resource real-world applications and the complexity of multi-modal processing. However, the vulnerability of LVLMs is relatively underexplored, posing potential security risks in daily usage. In this paper, we provide a comprehensive review of the various forms of existing LVLM attacks. Specifically, we first introduce the background of attacks targeting LVLMs, including the attack preliminary, attack challenges, and attack resources. Then, we systematically review the development of LVLM attack methods, such as adversarial attacks that manipulate model outputs, jailbreak attacks that exploit model vulnerabilities for unauthorized actions, prompt injection attacks that engineer the prompt type and pattern, and data poisoning that affects model training. Finally, we discuss promising research directions in the future. We believe that our survey provides insights into the current landscape of LVLM vulnerabilities, inspiring more researchers to explore and mitigate potential safety issues in LVLM developments. The latest papers on LVLM attacks are continuously collected in https://***/liudaizong/Awesome-LVLM-Attack. Copyright © 2024, The Authors. All rights reserved.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Expediting Distributed GNN Training with Feature-only Partition and Optimized Communication Planning

Expediting Distributed GNN Training with Feature-only Partit...

引用

IEEE Annual Joint Conference: INFOCOM, IEEE computer and Communications Societies

作者： Bingqian Du Jun Liu Ziyue Luo Chuan Wu Qiankun Zhang Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China Department of Electrical and Computer Engineering The Ohio State University USA Department of Computer Science The University of Hong Kong Hong Kong School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan China

ISBN: (数字)9798350383508

ISBN: (纸本)9798350383515

Feature-only partition of large graph data in distributed Graph Neural Network (GNN) training offers advantages over commonly adopted graph structure partition, such as minimal graph preprocessing cost and elimination of cross-worker subgraph sampling burdens. Nonetheless, performance bottleneck of GNN training with feature-only partitions still largely lies in the substantial communication overhead due to cross-worker feature fetching. To reduce the communication overhead and expedite distributed training, we first investigate and answer two key questions on convergence behaviors of GNN model in feature-partition based distribute GNN training: 1) As no worker holds a complete copy of each feature, can gradient exchange among workers compensate for the information loss due to incomplete local features? 2) If the answer to the first question is negative, is feature fetching in every training iteration of the GNN model necessary to ensure model convergence? Based on our theoretical findings on these questions, we derive an optimal communication plan that decides the frequency for feature fetching during the training process, taking into account bandwidth levels among workers and striking a balance between model loss and training time. Extensive evaluation demonstrates consistent results with our theoretical analysis, and the effectiveness of our proposed design.

关键词： Training Time-frequency analysis Costs Computational modeling Distributed databases Bandwidth Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Towards Context-aware Support for Color Vision Deficiency: An Approach Integrating LLM and AR

Towards Context-aware Support for Color Vision Deficiency: A...

引用

IEEE Global Conference on Consumer Electronics (GCCE)

作者： Shogo Morita Yan Zhang Takuto Yamauchi Sinan Chen Jialong Li Kenji Tei School of Computing Tokyo Institute of Technology Tokyo Japan International Research Center for Neurointelligence University of Tokyo Tokyo Japan Department of Computer Science and Engineering Waseda University Tokyo Japan Center of Mathematical and Data Sciences Kobe University Kobe Japan

ISBN: (数字)9798350355079

ISBN: (纸本)9798350355086

People with color vision deficiency often face challenges in distinguishing colors such as red and green, which can complicate daily tasks and require the use of assistive tools or environmental adjustments. Current support tools mainly focus on presentation-based aids, like the color vision modes found in iPhone accessibility settings. However, offering context-aware support, like indicating the doneness of meat, remains a challenge since task-specific solutions are not cost-effective for all possible scenarios. To address this, our paper proposes an application that provides contextual and autonomous assistance. This application is mainly composed of: (i) an augmented reality interface that efficiently captures context; and (ii) a multi-modal large language model-based reasoner that serves to cognitize the context and then reason about the appropriate support contents. Preliminary user experiments with two color vision deficient users across five different scenarios have demonstrated the effectiveness and universality of our application.

关键词： Green products Color Faces Consumer electronics Augmented reality Context modeling

来源：评论

学校读者我要写书评

暂无评论

A Trust-Based Personalized Differential Privacy Guarantees for Online Social Networks

引用

IEEE Transactions on Network and Service Management 2025年

作者： Chen, Jiajun Hu, Chunqiang Sheng, Weihong Li, Ruinian Zhao, Ruifeng Yu, Jiguo Chongqing University College of Computer Science Chongqing401331 China Chongqing University School of Big Data and Software Engineering Chongqing401331 China Bowling Green State University Computer Science Department United States Electric Power Dispatching and Control Center of Guangdong Power Grid Co. Ltd Guangzhou China University of Electronic Science and Technology of China School of Information and Software Engineering Chengdu611731 China Qilu University of Technology Big Data Institute Jinan250353 China

Online social networks have emerged as a significant data source, but the extensive collection and utilization of personal information have given rise to profound concerns regarding privacy. From a legislative and policy perspective, and in alignment with the concept of privacy as control, users have the right to control their personal privacy information. However, users often encounter challenges in terms of understanding and effectively managing their privacy settings to align with their specific privacy requirements. To address this issue, in this paper, we incorporate the concept of trust and propose a trust-based personalized differential privacy model for online social networks, denoted as TPDP, which relies on a trusted central server to facilitate its operation. Specifically, when a user requests access to another user's personal information, the TPDP mechanism provides a privacy response, where the privacy level is determined based on the direct and indirect trust values among users, calculated automatically by the trusted central server. Furthermore, the proposed TPDP model offers user-to-user personalized differential privacy protection from the perspectives of network structures, trust-related factors, and trust propagation patterns. Finally, we validate the model's feasibility and assess the privacy-utility trade-off, as well as its robustness against attacks, through theoretical analysis and performance evaluation. © 2004-2012 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Multi-Modal Knowledge Graph Transformer Framework for Multi-Modal Entity Alignment

arXiv

引用

arXiv 2023年

作者： Li, Qian Ji, Cheng Guo, Shu Liang, Zhaoji Wang, Lihong Li, Jianxin School of Computer Science and Engineering Beihang University Beijing China Beijing Advanced Innovation Center for Big Data and Brain Computing Beijing China National Computer Network Emergency Response Technical Team Coordination Center of China China

Multi-Modal Entity Alignment (MMEA) is a critical task that aims to identify equivalent entity pairs across multi-modal knowledge graphs (MMKGs). However, this task faces challenges due to the presence of different types of information, including neighboring entities, multimodal attributes, and entity types. Directly incorporating the above information (e.g., concatenation or attention) can lead to an unaligned information space. To address these challenges, we propose a novel MMEA transformer, called MoAlign, that hierarchically introduces neighbor features, multi-modal attributes, and entity types to enhance the alignment task. Taking advantage of the transformer's ability to better integrate multiple information, we design a hierarchical modifiable self-attention block in a transformer encoder to preserve the unique semantics of different information. Furthermore, we design two entity-type prefix injection methods to integrate entity-type information using type prefixes, which help to restrict the global information of entities not present in the MMKGs. Our extensive experiments on benchmark datasets demonstrate that our approach outperforms strong competitors and achieves excellent entity alignment performance. Copyright © 2023, The Authors. All rights reserved.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Detecting Backdoors During the Inference Stage Based on Corruption Robustness Consistency

Detecting Backdoors During the Inference Stage Based on Corr...

引用

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Xiaogeng Liu Minghui Li Haoyu Wang Shengshan Hu Dengpan Ye Hai Jin Libing Wu Chaowei Xiao School of Cyber Science and Engineering Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Software Engineering Huazhong University of Science and Technology School of Cyber Science and Engineering Wuhan University School of Computer Science and Technology Huazhong University of Science and Technology Cluster and Grid Computing Lab Arizona State University

Deep neural networks are proven to be vulnerable to backdoor attacks. Detecting the trigger samples during the inference stage, i.e., the test-time trigger sample detection, can prevent the backdoor from being triggered. However, existing detection methods often require the defenders to have high accessibility to victim models, extra clean data, or knowledge about the appearance of backdoor triggers, limiting their practicality. In this paper, we propose the test-time corruption robustness consistency evaluation (TeCo) 1 1 https://***/CGCL-codes/TeCo, a novel test-time trigger sample detection method that only needs the hard-label outputs of the victim models without any extra information. Our journey begins with the intriguing observation that the backdoor-infected models have similar performance across different image corruptions for the clean images, but perform discrepantly for the trigger samples. Based on this phenomenon, we design TeCo to evaluate test-time robustness consistency by calculating the deviation of severity that leads to predictions' transition across different corruptions. Extensive experiments demonstrate that compared with state-of-the-art defenses, which even require either certain information about the trigger types or accessibility of clean data, TeCo outperforms them on different backdoor attacks, datasets, and model architectures, enjoying a higher AUROC by 10% and 5 times of stability.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Using Lock-Free Design for Throughput-Optimized Cache Eviction 25

Using Lock-Free Design for Throughput-Optimized Cache Evicti...

引用

Abstracts of the 2025 ACM SIGMETRICS International Conference on Measurement and Modeling of computer Systems

作者： Chao Dong Fang Wang Hong Jiang Dan Feng Key Laboratory of Information Storage System (Wuhan National Laboratory for Optoelectronics) Engineering Research Center of data storage systems and Technology (School of Computer Science and Technology) Ministry of Education of China Huazhong University of Science and Techonolgy Wuhan Hubei China Department of Computer Science and Engineering University of Texas at Arlington Arlington Texas USA

ISBN: (纸本)9798400715938

This paper presents a practical approach to cache eviction algorithm design, called Mobius, that optimizes the concurrent throughput of caches and reduces cache operation latency by utilizing lock-free data structures, while maintaining high cache hit ratios. Mobius includes two key designs. First, Mobius employs two lock-free FIFO queues to manage cache items, ensuring that all cache operations are executed efficiently in concurrency. Second, Mobius integrates a consecutive detection mechanism that merges multiple modifications during eviction into a single operation, thereby reducing data races. The implementation of Mobius in CacheLib and RocksDB highlights its high concurrency in both synthetic and real-world workloads.

关键词： cache eviction policy

来源：评论

学校读者我要写书评

暂无评论

Downstream-agnostic Adversarial Examples

Downstream-agnostic Adversarial Examples

引用

International Conference on computer Vision (ICCV)

作者： Ziqi Zhou Shengshan Hu Ruizhi Zhao Qian Wang Leo Yu Zhang Junhui Hou Hai Jin School of Cyber Science and Engineering Huazhong University of Science and Technology National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Wuhan University School of Information and Communication Technology Griffith University Department of Computer Science City University of Hong Kong School of Computer Science and Technology Huazhong University of Science and Technology Cluster and Grid Computing Lab

Self-supervised learning usually uses a large amount of unlabeled data to pre-train an encoder which can be used as a general-purpose feature extractor, such that downstream users only need to perform fine-tuning operations to enjoy the benefit of "large model". Despite this promising prospect, the security of pre-trained encoder has not been thoroughly investigated yet, especially when the pre-trained encoder is publicly available for commercial *** this paper, we propose AdvEncoder, the first framework for generating downstream-agnostic universal adversarial examples based on the pre-trained encoder. AdvEncoder aims to construct a universal adversarial perturbation or patch for a set of natural images that can fool all the downstream tasks inheriting the victim pre-trained encoder. Unlike traditional adversarial example works, the pre-trained encoder only outputs feature vectors rather than classification labels. Therefore, we first exploit the high frequency component information of the image to guide the generation of adversarial examples. Then we design a generative attack framework to construct adversarial perturbations/patches by learning the distribution of the attack surrogate dataset to improve their attack success rates and transferability. Our results show that an attacker can successfully attack downstream tasks without knowing either the pre-training dataset or the downstream dataset. We also tailor four defenses for pre-trained encoders, the results of which further prove the attack ability of AdvEncoder. Our codes are available at: https://***/CGCL-codes/AdvEncoder.

关键词：

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：