Can adversarially robust learning leverage computational hardness?

MaRDI QIDQ7034382zbMATH OpenFDO

Authors Saeed Mahloujifar, Mohammad Mahmoody

Publication date 31 January 2025

Full work available at URL https://proceedings.mlr.press/v98/mahloujifar19a.html

zbMATH Keywords

adversarial examplesadversarial machine learningcomputational hardnesscryptographyevasion attackspoisoning attacksrobust learning

Mathematics Subject Classification ID

Computational difficulty of problems (lower bounds, completeness, difficulty of approximation, etc.) (68Q17) Learning and adaptive systems in artificial intelligence (68T05) Cryptography (94A60)

Cites work

• Adaptively secure coin-flipping, revisited
• Agnostic estimation of mean and covariance
• Asymptotic theory of finite dimensional normed spaces. With an appendix by M. Gromov: Isoperimetric inequalities in Riemannian manifolds
• Blockwise p-tampering attacks on cryptographic primitives, extractors, and learners
• Concentration of measure and isoperimetric inequalities in product spaces
• Efficient noise-tolerant learning from statistical queries
• scientific article; zbMATH DE number 4170917 (Why is no real title available?)
• scientific article; zbMATH DE number 1775450 (Why is no real title available?)
• scientific article; zbMATH DE number 7014202 (Why is no real title available?)
• Learning from untrusted data
• Learning in the Presence of Malicious Errors
• Learning under p-tampering attacks
• List-decodable robust mean estimation and learning mixtures of spherical Gaussians
• Optimal data-dependent hashing for approximate near neighbors
• PAC learning with nasty noise.
• Robust estimators in high dimensions without the computational intractability
• Robust optimization
• Robust probabilistic inference
• Some extremal problems arising from discrete control processes
• Statistical query lower bounds for robust estimation of high-dimensional Gaussians and Gaussian mixtures
• The power of localization for efficiently learning linear separators with noise
• Unconditional and symmetric sets in \(n\)-dimensional normed spaces

Cited in

(3)

• Adversarially robust learning could leverage computational hardness.
• A tight lower bound on adaptively secure full-information coin flip
• New algorithmic directions in optimal transport and applications for product spaces

This page was built for publication: Can adversarially robust learning leverage computational hardness?

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7034382)