Stronger data poisoning attacks break data sanitization defenses

DOI10.1007/S10994-021-06119-YMaRDI QIDQ2127214zbMATH OpenOpenAlexFDO

Authors Pang Wei Koh, Jacob Steinhardt, Percy Liang

Publication date 20 April 2022

Published in Machine Learning (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1811.00741

zbMATH Keywords

anomaly detectiondata poisoningdata sanitizationsecurity

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Cites work

• A survey of outlier detection methodologies
• Adaptive subgradient methods for online learning and stochastic optimization
• Hardness of learning halfspaces with noise
• Learning from untrusted data
• Learning geometric concepts with nasty noise
• Learning halfspaces with malicious noise
• Learning in the Presence of Malicious Errors
• Notes about the Carathéodory number
• On agnostic learning of parities, monomials, and halfspaces
• On the learnability and design of output codes for multiclass problems
• Practical bilevel optimization. Algorithms and applications
• Resilience: a criterion for learning in the presence of arbitrary outliers
• Robust estimators in high-dimensions without the computational intractability
• Second-order stochastic optimization for machine learning in linear time
• Security analysis of online centroid anomaly detection
• Some properties of the bilevel programming problem
• The security of machine learning

Cited in

(2)

• scientific article; zbMATH DE number 7626725 (Why is no real title available?)
• Random Projection and Recovery for High Dimensional Optimization with Arbitrary Outliers

Describes a project that uses

Uses Software

• IMDB
• AdaGrad
• BadNets
• Sever
• DeepFool

This page was built for publication: Stronger data poisoning attacks break data sanitization defenses

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2127214)