基于对抗鲁棒记忆的持续学习器
Adversarial Robust Memory-Based Continual Learner
AI总结：
针对持续学习对抗鲁棒性不足问题，提出基于记忆的持续学习器，通过调整logits和梯度数据选择机制缓解遗忘与梯度混淆，在多个基准上提升对抗准确率。
摘要：
尽管持续学习已取得显著进展，但此类方法的对抗脆弱性尚未得到充分讨论。我们深入研究了基于记忆的持续学习算法的对抗鲁棒性，并观察到直接应用对抗训练技术所带来的鲁棒性提升有限。初步研究揭示了构建对抗鲁棒持续学习器所面临的双重挑战：持续学习中的加速遗忘以及对抗鲁棒性中的梯度混淆。在本研究中，我们提出了一种新颖的基于对抗鲁棒记忆的持续学习器，该学习器通过调整数据logits来缓解由对抗样本引起的对过去知识的遗忘。此外，我们设计了一种基于梯度的数据选择机制，以克服由于存储数据有限而导致的梯度混淆。所提出的方法能够以即插即用的方式广泛集成到现有的基于记忆的持续学习以及对抗训练算法中。在Split-CIFAR10/100和Split-Tiny-ImageNet上的大量实验证明了我们方法的有效性，在对抗数据上实现了高达8.13%的准确率提升。
英文摘要：
Despite the remarkable advances that have been made in continual learning, the adversarial vulnerability of such methods has not been fully discussed. We delve into the adversarial robustness of memory-based continual learning algorithms and observe limited robustness improvement by directly applying adversarial training techniques. Preliminary studies reveal the twin challenges for building adversarial robust continual learners: accelerated forgetting in continual learning and gradient obfuscation in adversarial robustness. In this study, we put forward a novel adversarial robust memory-based continual learner that adjusts data logits to mitigate the forgetting of pasts caused by adversarial samples. Furthermore, we devise a gradient-based data selection mechanism to overcome the gradient obfuscation caused by limited stored data. The proposed approach can widely integrate with existing memory-based continual learning as well as adversarial training algorithms in a plug-and-play way. Extensive experiments on Split-CIFAR10/100 and Split-Tiny-ImageNet demonstrate the effectiveness of our approach, achieving up to 8.13% higher accuracy for adversarial data.
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。
