arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.01575cs.LG
收藏

通过潜在对抗扰动的可靠快速对抗训练

Reliably fast adversarial training via latent adversarial perturbation

KAIST(韩国科学技术院)

Geon Yeong Park, Sang Wan Lee

摘要:

虽然多步对抗训练作为对抗强对抗攻击的有效防御方法而广泛流行,但与标准训练相比,其计算成本非常昂贵。已经提出了几种单步对抗训练方法来减轻上述开销成本;然而,根据优化设置,它们的性能并不足够可靠。为了克服这些限制,我们偏离了现有的基于输入空间的对抗训练机制,并提出了一种单步潜在对抗训练方法(SLAT),该方法利用潜在表示的梯度作为潜在对抗扰动。我们证明了特征梯度的L1范数通过所采用的潜在扰动被隐式正则化,从而恢复局部线性并确保可靠的性能,与现有的单步对抗训练方法相比。由于潜在扰动是基于潜在表示的梯度,而这些梯度可以在输入梯度计算过程中免费获得,因此所提出的方法与快速梯度符号方法的时间大致相同。实验结果表明,尽管该方法结构简单,但性能优于最先进的加速对抗训练方法。

英文摘要:

While multi-step adversarial training is widely popular as an effective defense method against strong adversarial attacks, its computational cost is notoriously expensive, compared to standard training. Several single-step adversarial training methods have been proposed to mitigate the above-mentioned overhead cost; however, their performance is not sufficiently reliable depending on the optimization setting. To overcome such limitations, we deviate from the existing input-space-based adversarial training regime and propose a single-step latent adversarial training method (SLAT), which leverages the gradients of latent representation as the latent adversarial perturbation. We demonstrate that the L1 norm of feature gradients is implicitly regularized through the adopted latent perturbation, thereby recovering local linearity and ensuring reliable performance, compared to the existing single-step adversarial training methods. Because latent perturbation is based on the gradients of the latent representations which can be obtained for free in the process of input gradients computation, the proposed method costs roughly the same time as the fast gradient sign method. Experiment results demonstrate that the proposed method, despite its structural simplicity, outperforms state-of-the-art accelerated adversarial training methods.

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。

↑