arXiv 2506.078042025-06-09cs.LGcs.AIstat.ML

利用 Conformal Prediction 增强对抗鲁棒性：一个保证模型可靠性的框架

Enhancing Adversarial Robustness with Conformal Prediction: A Framework for Guaranteed Model Reliability

Huaiyin Institute of Technology, Huai'an, China（淮阴职业技术学院）；Chengdu Research Institute, City University of Hong Kong, Chengdu, China（香港城市大学成都研究院）；Saarland University, Saarbrücken, Germany（萨尔兰大学）；Alpha Benito Research, Los Angeles, USA（Alpha Benito研究公司）；City University of Hong Kong, Hong Kong, China（香港城市大学）

Jie Bao, Chuangyin Dang, Rui Luo, Hanwei Zhang, Zhixin Zhou

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出 OPSA 攻击及融合该攻击的 OPSA-AT conformal 对抗训练框架，在增大预测不确定性的同时提升模型对多种对抗攻击的鲁棒性与可靠性。

摘要：

随着深度学习模型越来越多地部署于高风险应用，针对对抗攻击的鲁棒防御以及可靠的性能保证变得至关重要。此外，仅凭准确率无法为这些模型提供充分保障或可靠的不确定性估计。本研究借助 Conformal Prediction（保形预测）原理推进对抗训练。具体而言，我们开发了一种名为 OPSA（OPtimal Size Attack，最优规模攻击）的对抗攻击方法，旨在通过最大化模型不确定性，在任意显著性水平下降低 conformal prediction 的效率，且不需要覆盖保证。相应地，我们提出 OPSA-AT（Adversarial Training，对抗训练）这一防御策略，将 OPSA 整合到一种新的 conformal 训练范式中。实验评估表明，与面向多种防御的基线方法相比，我们的 OPSA 攻击方法会引发更大的不确定性。相反，我们的 OPSA-AT 防御模型不仅显著增强了对 OPSA 的鲁棒性，也增强了对其他对抗攻击的鲁棒性，并保持可靠预测。我们的研究结果突显了这种整合方法在为安全关键领域开发可信且具韧性的深度学习模型方面的有效性。代码可在 https://github.com/bjbbbb/Enhancing-Adversarial-Robustness-with-Conformal-Prediction 获取。

英文摘要：

As deep learning models are increasingly deployed in high-risk applications, robust defenses against adversarial attacks and reliable performance guarantees become paramount. Moreover, accuracy alone does not provide sufficient assurance or reliable uncertainty estimates for these models. This study advances adversarial training by leveraging principles from Conformal Prediction. Specifically, we develop an adversarial attack method, termed OPSA (OPtimal Size Attack), designed to reduce the efficiency of conformal prediction at any significance level by maximizing model uncertainty without requiring coverage guarantees. Correspondingly, we introduce OPSA-AT (Adversarial Training), a defense strategy that integrates OPSA within a novel conformal training paradigm. Experimental evaluations demonstrate that our OPSA attack method induces greater uncertainty compared to baseline approaches for various defenses. Conversely, our OPSA-AT defensive model significantly enhances robustness not only against OPSA but also other adversarial attacks, and maintains reliable prediction. Our findings highlight the effectiveness of this integrated approach for developing trustworthy and resilient deep learning models for safety-critical domains. Our code is available at https://github.com/bjbbbb/Enhancing-Adversarial-Robustness-with-Conformal-Prediction.

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。