all AI news
Overcoming the Paradox of Certified Training with Gaussian Smoothing
March 13, 2024, 4:41 a.m. | Stefan Balauca, Mark Niklas M\"uller, Yuhao Mao, Maximilian Baader, Marc Fischer, Martin Vechev
cs.LG updates on arXiv.org arxiv.org
Abstract: Training neural networks with high certified accuracy against adversarial examples remains an open problem despite significant efforts. While certification methods can effectively leverage tight convex relaxations for bound computation, in training, these methods perform worse than looser relaxations. Prior work hypothesized that this is caused by the discontinuity and perturbation sensitivity of the loss surface induced by these tighter relaxations. In this work, we show theoretically that Gaussian Loss Smoothing can alleviate both of these …
abstract accuracy adversarial adversarial examples arxiv certification computation cs.lg examples networks neural networks paradox prior training type work
More from arxiv.org / cs.LG updates on arXiv.org
The Perception-Robustness Tradeoff in Deterministic Image Restoration
2 days, 2 hours ago |
arxiv.org
Jobs in AI, ML, Big Data
Founding AI Engineer, Agents
@ Occam AI | New York
AI Engineer Intern, Agents
@ Occam AI | US
AI Research Scientist
@ Vara | Berlin, Germany and Remote
Data Architect
@ University of Texas at Austin | Austin, TX
Data ETL Engineer
@ University of Texas at Austin | Austin, TX
Lead GNSS Data Scientist
@ Lurra Systems | Melbourne