Computational Limitations in Robust Classification and Win-Win Results

IACR Cryptology ePrint Archive, pp. 2182019.

Cited by: 0|Bibtex|Views46
EI
Other Links: dblp.uni-trier.de|academic.microsoft.com|arxiv.org

Abstract:

We continue the study of computational limitations in learning robust classifiers, following the recent work of Bubeck, Lee, Price and Razenshteyn. First, we demonstrate classification tasks where computationally efficient robust classifiers do not exist, even when computationally unbounded robust classifiers do. We rely on the hardness o...More

Code:

Data:

Full Text
Your rating :
0

 

Tags
Comments