arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2103.00381cs.LGstat.ML
收藏

对抗性信息瓶颈

Adversarial Information Bottleneck

Academy of Mathematics and Systems Science, Chinese Academy of Sciences(中国科学院数学与系统科学研究院);School of Mathematical Sciences, University of Chinese Academy of Sciences(中国科学院大学数学科学学院)

Penglong Zhai, Shihua Zhang

摘要:

信息瓶颈(IB)原理已经被用来解释深度学习的信息压缩和预测方面,这是平衡的权衡超参数。如何优化IB原则以提高鲁棒性和通过权衡超参数来计算压缩效果是两个具有挑战性的问题。先前的方法试图通过将随机噪声引入到学习表示中来优化IB原理,并且在讨厌的信息压缩和语义信息抽取中实现了最先进的性能。然而,它们在抵抗对抗性扰动方面的表现却远不如人意。为此,我们提出了一种对抗性信息瓶颈(AIB)方法,该方法没有任何关于表示的底层分布的明确假设,可以通过求解Min-Max优化问题来有效地优化。在合成和真实世界数据集上的数值实验表明,与几种竞争性IB方法相比,它在学习更多不变表示和减轻对抗性扰动方面是有效的。此外,我们还对比分析了各种IB方法的对抗鲁棒性,发现在IB曲线拐点处具有超参数$β$的IB模型在压缩和预测之间取得了最佳的折衷,对各种攻击具有最好的鲁棒性.

英文摘要:

The information bottleneck (IB) principle has been adopted to explain deep learning in terms of information compression and prediction, which are balanced by a trade-off hyperparameter. How to optimize the IB principle for better robustness and figure out the effects of compression through the trade-off hyperparameter are two challenging problems. Previous methods attempted to optimize the IB principle by introducing random noise into learning the representation and achieved state-of-the-art performance in the nuisance information compression and semantic information extraction. However, their performance on resisting adversarial perturbations is far less impressive. To this end, we propose an adversarial information bottleneck (AIB) method without any explicit assumptions about the underlying distribution of the representations, which can be optimized effectively by solving a Min-Max optimization problem. Numerical experiments on synthetic and real-world datasets demonstrate its effectiveness on learning more invariant representations and mitigating adversarial perturbations compared to several competing IB methods. In addition, we analyse the adversarial robustness of diverse IB methods contrasting with their IB curves, and reveal that IB models with the hyperparameter $β$ corresponding to the knee point in the IB curve achieve the best trade-off between compression and prediction, and has best robustness against various attacks.

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。

↑