首页/论文/2011.14045

arXiv 2011.140452020-11cs.LGcs.CR

Incorporating Hidden Layer representation into Adversarial Attacks and Defences

Shenzhen University（深圳大学）

Haojing Shen, Sihong Chen, Ran Wang, Xizhao Wang

英文摘要：

In this paper, we propose a defence strategy to improve adversarial robustness by incorporating hidden layer representation. The key of this defence strategy aims to compress or filter input information including adversarial perturbation. And this defence strategy can be regarded as an activation function which can be applied to any kind of neural network. We also prove theoretically the effectiveness of this defense strategy under certain conditions. Besides, incorporating hidden layer representation we propose three types of adversarial attacks to generate three types of adversarial examples, respectively. The experiments show that our defence method can significantly improve the adversarial robustness of deep neural networks which achieves the state-of-the-art performance even though we do not adopt adversarial training.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

阅读原文

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

本文内容

英文摘要相关论文

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

继续探索

相关论文

A simple defense against adversarial attacks on heatmap explanations

Technical University Denmark（丹麦技术大学）

With machine learning models being used for more sensitive applications, we rely on interpretability methods to prove that no discriminating attributes were used for classification. A potential concern is the so-called "fair-washing" - manipulating a model such that the features used in reality are hidden and more innocuous features are shown to be important instead. In our work we present an effective defence against such adversarial attacks on neural networks. By a simple aggregation of multiple explanation methods, the network becomes robust against manipulation. This holds even when the at…

2020-07-13cs.AIcs.LGstat.ML

Adversarial AI in Insurance: Pervasiveness and Resilience

University of Torino（都灵大学）；Collegio Carlo Alberto（卡洛·阿尔贝托学院）；Reale Mutua Group（瑞莱穆图阿集团）；KPA Group（KPA集团）；Samuel Neaman Institute（塞缪尔·尼曼研究所）；Technion（以色列理工学院）

The rapid and dynamic pace of Artificial Intelligence (AI) and Machine Learning (ML) is revolutionizing the insurance sector. AI offers significant, very much welcome advantages to insurance companies, and is fundamental to their customer-centricity strategy. It also poses challenges, in the project and implementation phase. Among those, we study Adversarial Attacks, which consist of the creation of modified input data to deceive an AI system and produce false outputs. We provide examples of attacks on insurance AI applications, categorize them, and argue on defence methods and precautionary s…

2023-01-17cs.LGq-fin.GN

第一道防线：稳健的第一层可缓解对抗攻击

First line of defense: A robust first layer mitigates adversarial attacks

Indian Institute of Technology Madras（马德拉斯印度理工学院）；Imperial College London（帝国理工学院）

本文提出通过结合大卷积核、增加卷积过滤器和最大池化操作构建隐式对抗噪声过滤器（ANF）作为神经网络第一层，无需对抗训练即可提升 ResNet 等架构的对抗鲁棒性，效果媲美对抗训练模型。

2024-08-21cs.LG

Incorporating Expressive Graphical Models in Variational Approximations: Chain-Graphs and Hidden Variables

School of Computer Science & Engineering, Hebrew University（希伯来大学计算机科学与工程学院）

Global variational approximation methods in graphical models allow efficient approximate inference of complex posterior distributions by using a simpler model. The choice of the approximating model determines a tradeoff between the complexity of the approximation procedure and the quality of the approximation. In this paper, we consider variational approximations based on two classes of models that are richer than standard Bayesian networks, Markov networks or mixture models. As such, these classes allow to find better tradeoffs in the spectrum of approximations. The first class of models are …

2013-01-10cs.AIcs.LG

解释无法解释的内容：揭示隐藏相关性以提升可解释性

Explaining the Unexplained: Revealing Hidden Correlations for Better Interpretability

Tamkang University（淡江大学）；Ming Chuan University（铭传大学）；National Institute of Technology, Shillong（希隆国家理工学院）

针对深度学习“黑箱”问题及现有可解释方法未充分考虑特征相关性、评估不足的局限，提出RealExp方法，通过解耦Shapley值量化单特征重要性与相关性重要性，提升解释可靠性，在图像分类和文本任务中优于现有方法并具实用价值。

2025-02-09cs.AIcs.LG

SemRoDe：通过宏观对抗训练学习对词级攻击鲁棒的表示

SemRoDe: Macro Adversarial Training to Learn Representations That are Robust to Word-Level Attacks

National University of Singapore（新加坡国立大学）；Institute for Infocomm Research（信息通信研究院）；A*STAR（新加坡科技研究局）；Centre for Frontier AI Research（前沿人工智能研究中心）；Nanyang Technological University（南洋理工大学）

本文提出SemRoDe宏观对抗训练方法，通过将对抗样本投影到偏移最小的域并对齐高层特征，学习桥接基础域与对抗域的鲁棒表示，在BERT和RoBERTa上取得最先进的词级攻击防御效果。

2024-03-27cs.CLcs.LG

更多相关论文收起更多论文

Factorial Hidden Markov Models for Learning Representations of Natural Language

Temple University（天普大学）

Most representation learning algorithms for language and image processing are local, in that they identify features for a data point based on surrounding points. Yet in language processing, the correct meaning of a word often depends on its global context. As a step toward incorporating global context into representation learning, we develop a representation learning algorithm that incorporates joint prediction into its technique for producing features for a word. We develop efficient variational methods for learning Factorial Hidden Markov Models from large texts, and use variational distribu…

2014-02-18cs.CLcs.LG

足够好的LLM混淆（GELO）

Good-Enough LLM Obfuscation (GELO)

SingularityNET Foundation（SingularityNET基金会）；Singularity Compute

提出GELO协议，通过每批可逆混合隐藏状态，在共享加速器上保护LLM推理隐私，抵抗统计攻击，计算开销约20-30%。

2026-06-23cs.CRcs.LG

对比推理对齐：从隐藏表示中学习强化学习

Contrastive Reasoning Alignment: Reinforcement Learning from Hidden Representations

Northwestern University（西北大学）；University of Michigan（密歇根大学）；Illinois Institute of Technology（伊利诺伊理工学院）

本文提出了一种基于对比学习和强化学习的框架CRAFT，通过优化隐藏状态空间中的目标来提升对抗攻击的鲁棒性，核心贡献是通过隐藏空间的几何结构实现推理层面的安全对齐。

2026-05-19cs.AIcs.CLcs.LG

Deep Global-Connected Net With The Generalized Multi-Piecewise ReLU Activation in Deep Learning

University of Waterloo（滑铁卢大学）

Recent Progress has shown that exploitation of hidden layer neurons in convolution neural networks incorporating with a carefully designed activation function can yield better classification results in the field of computer vision. The paper firstly introduces a novel deep learning architecture aiming to mitigate the gradient-vanishing problem, in which the earlier hidden layer neurons could be directly connected with the last hidden layer and feed into the last layer for classification. We then design a generalized linear rectifier function as the activation function that can approximate arbi…

2018-06-19cs.CVcs.LG

一种带区域分解的多可迁移神经网络方法用于椭圆界面问题

A Multiple Transferable Neural Network Method with Domain Decomposition for Elliptic Interface Problems

School of Mathematical Sciences, Beihang University（北京航空航天大学数学科学学院）；Department of Mathematics, University of South Carolina（南卡罗来纳大学数学系）

本文提出一种结合区域分解与界面条件的多可迁移神经网络方法，用于高效求解具有不连续解和导数的椭圆界面问题，并通过经验公式与自适应权重策略提升精度和鲁棒性。

2025-02-27cs.LGcs.NAmath.NA