Despite their numerous successes, there are many scenarios where adversarial risk metrics do not provide an appropriate measure of robustness. For example, test-time perturbations may occur in a probabilistic manner rather than being generated by an explicit adversary, while the poor train--test generalization of adversarial metrics can limit their usage to simple problems. Motivated by this, we develop a probabilistic robust risk framework, the statistically robust risk (SRR), which considers pointwise corruption distributions, as opposed to worst-case adversaries. The SRR provides a distinct and complementary measure of robust performance, compared to natural and adversarial risk. We show that the SRR admits estimation and training schemes which are as simple and efficient as for the natural risk: these simply require noising the inputs, but with a principled derivation for exactly how and why this should be done. Furthermore, we demonstrate both theoretically and experimentally that it can provide superior generalization performance compared with adversarial risks, enabling application to high-dimensional datasets.
翻译:尽管取得了许多成功,但有许多情况是,对抗性风险指标不能提供适当的稳健度量度,例如,测试时的扰动可能以概率方式发生,而不是由明确的对手产生,而对对抗性指标的测试失败的训练测试一般化可以将其使用限制在简单的问题上,因此,我们开发了一个概率稳健的风险框架,即统计上稳健的风险(SRR),它考虑到点性腐败分布,而不是最坏的对手。SRR提供了一种与自然风险和对抗性风险相比的稳健性表现的明显和互补的衡量标准。我们表明,SRR接受的估算和培训计划与自然风险一样简单和有效:这些只是需要对投入进行消毒,但有原则性地推断出应如何和为什么这样做。此外,我们从理论上和实验上都表明,它能够提供比对抗性风险更优的概括性业绩,从而能够应用于高度的数据集。