Study proposes representational geometry as a diagnostic for adversarial robustness
A new arXiv paper introduces representational geometry as a diagnostic tool for assessing how well neural networks resist adversarial perturbations, complementing adversarial training approaches.
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-11-16 05:00 · arXiv cs.LG
Representational Geometry as a Diagnostic for Adversarial Robustness