检索结果-内蒙古大学图书馆

Regularization Mixup Adversarial Training: A Defense Strategy for Membership Privacy with Model Availability Assurance 2

学校读者我要写书评

暂无评论

Regularization Mixup Adversarial Training: A Defense Strateg...

2nd International Conference on Big data and Privacy Computing, BDPC 2024

作者： Ding, Zehua Tian, Youliang Wang, Guorong Xiong, Jinbo College of Computer Science and Technology Guizhou University State Key Laboratory of Public Big Data Guiyang China College of Computer and Cyber Security Fujian Normal University Fujian Provincial Key Laboratory of Network Security and Cryptology Fuzhou China

ISBN: (纸本)9798350307597

Neural network models face two highly destructive threats in real-world applications: membership inference attacks (MIAs) and adversarial attacks (AAs). One compromises the model's confidentiality, leading to membership privacy breaches, while the other disrupts the model's availability, rendering it dysfunctional. Recent work has shown that adversarial examples can pose dual threats to neural network models, both MIAs and AAs, so that these two different types of attacks have a certain degree of intersection. However, existing defense methods typically focus on defending against one of the two attacks and cannot simultaneously address both. To address the above issues, we propose a defense method called regularization mixup adversarial training (RMAT), aiming to protect model membership privacy while ensuring model availability. The core idea of RMAT is to use a continuous augmentation strategy during training to generate mixup adversarial examples, weakening attackers' understanding of membership data, and updating the model with a new membership-weighted loss strategy to improve its generalization ability for handling unknown samples. This approach can further protect the model's membership privacy and ensure model availability. Experimental results demonstrate that compared to single defense strategies, RMAT can simultaneously protect the model's membership privacy and availability. © 2024 IEEE.

关键词： Adversarial machine learning

Reproducibility Analysis and Enhancements for Multi-Aspect Dense Retriever with Aspect Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Bi, Keping Sun, Xiaojie Guo, Jiafeng Cheng, Xueqi CAS Key Lab of Network Data Science and Technology ICT CAS China University of Chinese Academy of Sciences China

Multi-aspect dense retrieval aims to incorporate aspect information (e.g., brand and category) into dual encoders to facilitate relevance matching. As an early and representative multi-aspect dense retriever, MADRAL learns several extra aspect embeddings and fuses the explicit aspects with an implicit aspect "OTHER" for final representation. MADRAL was evaluated on proprietary data and its code was not released, making it challenging to validate its effectiveness on other datasets. We failed to reproduce its effectiveness on the public MA-Amazon data, motivating us to probe the reasons and re-examine its components. We propose several component alternatives for comparisons, including replacing "OTHER" with "CLS" and representing aspects with the first several content tokens. Through extensive experiments, we confirm that learning "OTHER" from scratch in aspect fusion is harmful. In contrast, our proposed variants can greatly enhance the retrieval performance. Our research not only sheds light on the limitations of MADRAL but also provides valuable insights for future studies on more powerful multi-aspect dense retrieval models. Code will be released at: https://***/sunxiaojie99/Reproducibility-for-MADRAL. © 2024, CC BY.

关键词：

Are Large Language Models More Honest in Their Probabilistic or Verbalized Confidence?

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ni, Shiyu Bi, Keping Yu, Lulu Guo, Jiafeng CAS Key Lab of Network Data Science and Technology ICT CAS China University of Chinese Academy of Sciences China

Large language models (LLMs) have been found to produce hallucinations when the question exceeds their internal knowledge boundaries. A reliable model should have a clear perception of its knowledge boundaries, providing correct answers within its scope and refusing to answer when it lacks knowledge. Existing research on LLMs’ perception of their knowledge boundaries typically uses either the probability of the generated tokens or the verbalized confidence as the model’s confidence in its response. However, these studies overlook the differences and connections between the two. In this paper, we conduct a comprehensive analysis and comparison of LLMs’ probabilistic perception and verbalized perception of their factual knowledge boundaries. First, we investigate the pros and cons of these two perceptions. Then, we study how they change under questions of varying frequencies. Finally, we measure the correlation between LLMs’ probabilistic confidence and verbalized confidence. Experimental results show that 1) LLMs’ probabilistic perception is generally more accurate than verbalized perception but requires an in-domain validation set to adjust the confidence threshold. 2) Both perceptions perform better on less frequent questions. 3) It is challenging for LLMs to accurately express their internal confidence in natural language. © 2024, CC BY.

关键词： Depth perception

A Novel 3D Medical Image Segmentation Model Using Improved SAM

学校读者我要写书评

暂无评论

A Novel 3D Medical Image Segmentation Model Using Improved S...

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Kuang, Yuansen Ma, Xitong Zhao, Jing Wang, Guangchen Zeng, Yijie Liu, Song Key Laboratory of Computing Power Network and Information Security Jinan China Shandong Engineering Research Center of Big Data Applied Technology Faculty of Computer Science and Technology Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9781665410205

3D medical image segmentation is an essential task in the medical image field, which aims to segment organs or tumours into different labels. A number of issues exist with the current 3D medical image segmentation task: existing models cannot simultaneously obtain the space correlation and depth correlation of 3D slices;previous models suffer from local detail loss of positional embedding in 3D images;previous approaches often have blurring of boundaries in segmenting 3D images. To solve these shortcomings, we propose a 3D medical image segmentation model named TPM-SAM. In our model, we design a twinchannel image encoder to simultaneously capture the space correlation and depth correlation of 3D slices through a multi-head attention mechanism and improved adapters. Furthermore, we design a prompt encoding generator, which divides the volumetric image into small blocks and better captures the local detail information. In addition, we introduce a multi-layer aggregation decoder by employing U-Net with multi-level skip connection to solve the blurring of boundaries in processing 3D images. Finally, we experimented and evaluated our model on KiTS21 and LiTS17 datasets to compare with other baseline models. © 2024 IEEE.

关键词： Image segmentation

Evaluating Implicit Bias in Large Language Models by Attacking From a Psychometric Perspective

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wen, Yuchen Bi, Keping Chen, Wei Guo, Jiafeng Cheng, Xueqi CAS Key Laboratory of AI Safety ICT CAS Beijing China CAS Key Lab of Network Data Science and Technology ICT CAS Beijing China University of Chinese Academy of Sciences Beijing China

As large language models (LLMs) become an important way of information access, there have been increasing concerns that LLMs may intensify the spread of unethical content, including implicit bias that hurts certain populations without explicit harmful words. In this paper, we conduct a rigorous evaluation of LLMs’ implicit bias towards certain demographics by attacking them from a psychometric perspective to elicit agreements to biased viewpoints. Inspired by psychometric principles in cognitive and social psychology, we propose three attack approaches, i.e., Disguise, Deception, and Teaching. Incorporating the corresponding attack instructions, we built two benchmarks: (1) a bilingual dataset with biased statements covering four bias types (2.7K instances) for extensive comparative analysis, and (2) BUMBLE, a larger benchmark spanning nine common bias types (12.7K instances) for comprehensive evaluation. Extensive evaluation of popular commercial and open-source LLMs shows that our methods can elicit LLMs’ inner bias more effectively than competitive baselines. Our attack methodology and benchmarks offer an effective means of assessing the ethical risks of LLMs, driving progress toward greater accountability in their development. * © 2024, CC BY-NC-SA.

关键词： Risk assessment

Scale‐wise interaction fusion and knowledge distillation network for aerial scene recognition

学校读者我要写书评

暂无评论

CAAI Transactions on Intelligence technology 2023年第4期8卷 1178-1190页

作者： Hailong Ning Tao Lei Mengyuan An Hao Sun Zhanxuan Hu Asoke K.Nandi School of Computer Science and Technology Xi'an University of Posts and TelecommunicationsShaanxi Key Laboratory of Network Data Analysis and Intelligent ProcessingXi'anChina Xi'an Key Laboratory of Big Data and Intelligent Computing Xi'anChina School of Electronic Information and Artificial Intelligence Shaanxi University of Science and TechnologyXi'anChina School of Computer Central China Normal UniversityWuhanChina Department of Electronic and Electrical Engineering Brunel University LondonLondonUK Xi'an Jiaotong University Xi'anChina

Aerial scene recognition(ASR)has attracted great attention due to its increasingly essential *** of the ASR methods adopt the multi‐scale architecture because both global and local features play great roles in ***,the existing multi‐scale methods neglect the effective interactions among different scales and various spatial locations when fusing global and local features,leading to a limited ability to deal with challenges of large‐scale variation and complex background in aerial scene *** addition,existing methods may suffer from poor generalisations due to millions of to‐belearnt parameters and inconsistent predictions between global and local *** tackle these problems,this study proposes a scale‐wise interaction fusion and knowledge distillation(SIF‐KD)network for learning robust and discriminative features with scaleinvariance and background‐independent *** main highlights of this study include two *** the one hand,a global‐local features collaborative learning scheme is devised for extracting scale‐invariance features so as to tackle the large‐scale variation problem in aerial scene ***,a plug‐and‐play multi‐scale context attention fusion module is proposed for collaboratively fusing the context information between global and local *** the other hand,a scale‐wise knowledge distillation scheme is proposed to produce more consistent predictions by distilling the predictive distribution between different scales during *** experimental results show the proposed SIF‐KD network achieves the best overall accuracy with 99.68%,98.74%and 95.47%on the UCM,AID and NWPU‐RESISC45 datasets,respectively,compared with state of the arts.

关键词： deep learning image analysis image classification information fusion

CausalDiff: Causality-Inspired Disentanglement via Diffusion Model for Adversarial Defense

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhang, Mingkun Bi, Keping Chen, Wei Chen, Quanrun Guo, Jiafeng Cheng, Xueqi CAS Key Laboratory of AI Safety Institute of Computing Technology CAS China Key Laboratory of Network Data Science and Technology Institute of Computing Technology CAS China School of Statistics University of International Business and Economics China CAS stands for Chinese Academy of Sciences China

Despite ongoing efforts to defend neural classifiers from adversarial attacks, they remain vulnerable, especially to unseen attacks. In contrast, humans are difficult to be cheated by subtle manipulations, since we make judgments only based on essential factors. Inspired by this observation, we attempt to model label generation with essential label-causative factors and incorporate label-non-causative factors to assist data generation. For an adversarial example, we aim to discriminate the perturbations as non-causative factors and make predictions only based on the label-causative factors. Concretely, we propose a casual diffusion model (CausalDiff) that adapts diffusion models for conditional data generation and disentangles the two types of casual factors by learning towards a novel casual information bottleneck objective. Empirically, CausalDiff has significantly outperformed state-of-the-art defense methods on various unseen attacks, achieving an average robustness of 86.39% (+4.01%) on CIFAR-10, 56.25% (+3.13%) on CIFAR-100, and 82.62% (+4.93%) on GTSRB (German Traffic Sign Recognition Benchmark). The code is available at https://***/TMLResearchGroup-cas/CausalDiff. Copyright © 2024, The Authors. All rights reserved.

关键词： Traffic signs

ESNet: Evolution and Succession network for High-Resolution Salient Object Detection 41

学校读者我要写书评

暂无评论

ESNet: Evolution and Succession Network for High-Resolution ...

41st International Conference on Machine Learning, ICML 2024

作者： Liu, Hongyu Cong, Runmin Li, Hua Xu, Qianqian Huang, Qingming Zhang, Wei Institute of Information Science Beijing Jiaotong University Beijing Key Laboratory of Advanced Information Science and Network Technology Beijing China School of Control Science and Engineering Shandong University Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan China School of Computer Science and Technology Hainan University Hainan China Key Laboratory of Intelligent Information Processing Institute of Computing Technology CAS Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

Preserving details and avoiding high computational costs are the two main challenges for the High-Resolution Salient Object Detection (HRSOD) task. In this paper, we propose a two-stage HRSOD model from the perspective of evolution and succession, including an evolution stage with Low-resolution Location Model (LrLM) and a succession stage with High-resolution Refinement Model (HrRM). The evolution stage achieves detail-preserving salient objects localization on the low-resolution image through the evolution mechanisms on supervision and feature;the succession stage utilizes the shallow high-resolution features to complement and enhance the features inherited from the first stage in a lightweight manner and generate the final high-resolution saliency prediction. Besides, a new metric named Boundary-Detail-aware Mean Absolute Error (MAEBD) is designed to evaluate the ability to detect details in high-resolution scenes. Extensive experiments on five datasets demonstrate that our network achieves superior performance at real-time speed (49 FPS) compared to state-of-the-art methods. Our code is publicly available at: https://***/rmcong/ESNet_ICML24. Copyright 2024 by the author(s)

关键词： Miocene

Research on active defense decision-making method for cloud boundary networks based on reinforcement learning of intelligent agent

学校读者我要写书评

暂无评论

High-Confidence Computing 2024年第2期4卷 50-61页

作者： Huan Wang Yunlong Tang Yan Wang Ning Wei Junyi Deng Zhiyan Bin Weilong Li School of Computer Science and Technology Guangxi University of Science and TechnologyLiuzhou 545000China Liuzhou Key Laboratory of Big Data Intelligent Processing and Security Liuzhou 545000China Guangxi Education System Network Security Monitoring Center Liuzhou 545000China Center of Graphics and Information Guangxi ECO-Engineering Vocational&Technical CollegeLiuzhou 545004China

The cloud boundary network environment is characterized by a passive defense strategy,discrete defense actions,and delayed defense feedback in the face of network attacks,ignoring the influence of the external environment on defense decisions,thus resulting in poor defense ***,this paper proposes a cloud boundary network active defense model and decision method based on the reinforcement learning of intelligent agent,designs the network structure of the intelligent agent attack and defense game,and depicts the attack and defense game process of cloud boundary network;constructs the observation space and action space of reinforcement learning of intelligent agent in the non-complete information environment,and portrays the interaction process between intelligent agent and environment;establishes the reward mechanism based on the attack and defense gain,and encourage intelligent agents to learn more effective defense *** designed active defense decision intelligent agent based on deep reinforcement learning can solve the problems of border dynamics,interaction lag,and control dispersion in the defense decision process of cloud boundary networks,and improve the autonomy and continuity of defense decisions.

关键词： Active defense decision-making Cloud boundary network security Intelligent agent reinforcement learning Offensive and defensive game