Feature purification: how adversarial training performs robust deep learning

DOI10.1109/FOCS52979.2021.00098MaRDI QIDQ6943452zbMATH OpenFDO

Authors Zeyuan Allen Zhu, Yuanzhi Li

Publication date 13 August 2025

Full work available at URL https://doi.org/10.1109/FOCS52979.2021.00098

Mathematics Subject Classification ID

Proceedings, conferences, collections, etc. pertaining to computer science (68-06) Proceedings, conferences, collections, etc. pertaining to information and communication theory (94-06)

Cited in

(2)

• Random pruning over-parameterized neural networks can improve generalization: a training dynamics analysis
• High-dimensional regression with missing data: an asymptotic study

This page was built for publication: Feature purification: how adversarial training performs robust deep learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6943452)