A Finer Calibration Analysis for Adversarial Robustness
This work addresses calibration issues in adversarial robustness for machine learning practitioners, but it is incremental as it builds on and refines existing theoretical frameworks.
The paper tackles the problem of calibration analysis for adversarially robust classification by introducing a finer definition that generalizes previous work to cover most common hypothesis sets, fixing and extending prior results.
We present a more general analysis of $H$-calibration for adversarially robust classification. By adopting a finer definition of calibration, we can cover settings beyond the restricted hypothesis sets studied in previous work. In particular, our results hold for most common hypothesis sets used in machine learning. We both fix some previous calibration results (Bao et al., 2020) and generalize others (Awasthi et al., 2021). Moreover, our calibration results, combined with the previous study of consistency by Awasthi et al. (2021), also lead to more general $H$-consistency results covering common hypothesis sets.