pith. sign in

arxiv: 2204.00487 · v1 · pith:Z7O3ZODEnew · submitted 2022-04-01 · 💻 cs.LG · cs.AI· cs.CR

Robust and Accurate -- Compositional Architectures for Randomized Smoothing

classification 💻 cs.LG cs.AIcs.CR
keywords accuracystandardaccurateacesapproachescertifiablychallengingcompositional
0
0 comments X
read the original abstract

Randomized Smoothing (RS) is considered the state-of-the-art approach to obtain certifiably robust models for challenging tasks. However, current RS approaches drastically decrease standard accuracy on unperturbed data, severely limiting their real-world utility. To address this limitation, we propose a compositional architecture, ACES, which certifiably decides on a per-sample basis whether to use a smoothed model yielding predictions with guarantees or a more accurate standard model without guarantees. This, in contrast to prior approaches, enables both high standard accuracies and significant provable robustness. On challenging tasks such as ImageNet, we obtain, e.g., $80.0\%$ natural accuracy and $28.2\%$ certifiable accuracy against $\ell_2$ perturbations with $r=1.0$. We release our code and models at https://github.com/eth-sri/aces.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.