Adversarial Robustness as a Prior for Learned Representations
Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:AEBDDFSRrecord.jsonopen to challenge →
read the original abstract
An important goal in deep learning is to learn versatile, high-level feature representations of input data. However, standard networks' representations seem to possess shortcomings that, as we illustrate, prevent them from fully realizing this goal. In this work, we show that robust optimization can be re-cast as a tool for enforcing priors on the features learned by deep neural networks. It turns out that representations learned by robust models address the aforementioned shortcomings and make significant progress towards learning a high-level encoding of inputs. In particular, these representations are approximately invertible, while allowing for direct visualization and manipulation of salient input features. More broadly, our results indicate adversarial robustness as a promising avenue for improving learned representations. Our code and models for reproducing these results is available at https://git.io/robust-reps .
This paper has not been read by Pith yet.
Forward citations
Cited by 5 Pith papers
-
Toy Models of Superposition
Toy models demonstrate that polysemanticity arises when neural networks store more sparse features than neurons via superposition, producing a phase transition tied to polytope geometry and increased adversarial vulne...
-
Beyond False Stability: High-Noise Drift Gating for Test-Time Adversarial Defenses in Vision-Language Models
High-noise feature drift distinguishes adversarial from clean inputs in CLIP, allowing a plug-in gating mechanism to selectively trigger existing test-time defenses and raise mean clean+adversarial accuracy across 13 ...
-
Adjoint Inversion Reveals Holographic Superposition and Destructive Interference in CNN Classifiers
CNN classifiers work by holographic superposition and destructive interference in pixel space rather than selecting cleaned features, as proven by a new adjoint inversion framework that also yields a covariance-volume...
-
Toward Understanding Adversarial Distillation: Why Robust Teachers Fail
Adversarial distillation improves student robustness when teachers show high uncertainty on robustly unlearnable samples, suppressing noise memorization and allowing reliance on learnable robust signals.
-
Laundering AI Authority with Adversarial Examples
Adversarial examples enable AI authority laundering by causing production VLMs to give authoritative but wrong responses on subtly perturbed images, with success rates of 22-100% using decade-old attack methods.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.