This work develops a method for optimizing the auditor's canary set to improve privacy auditing, leveraging recent work on metagradient optimization and demonstrates that in certain instances, using such optimized canaries can improve empirical lower bounds for differentially private image classification models by seve...
Matteo Boglioni, Terrance Liu, Andrew Ilyas et al.· arXiv.org· 7 citations· ⚡1
Two algorithms, MAGE and SPELL, suited for reconstruction and attribution respectively, are presented, that run on existing metagradient machinery at no extra cost and demonstrate strong performance over existing baselines across training scales and measurement budgets.
Yu-Xi Chen, Hamza Golubovic, Han Tong et al.· 0 citations
This work formalizes poison selection as oracle-budgeted set optimization and introduces SAILS (Set-level Audit-Informed Iterative Learned Selection), which learns a set scorer from a few hundred finetune-and-evaluate runs, ranks millions of candidate sets, and audits only a small shortlist.
Aashiq Muhamed, Mona T. Diab, Virginia Smith et al.· 0 citations
It is found that planting harmful but benign-sounding reasoning in the actor's context can steer it to perform adversarial actions while evading monitors, an attack the authors term"plan injection".
Keertana Chidambaram, Andrew Ilyas, Vasilis Syrgkanis· 0 citations
This work introduces a framework for black-box audits of personalization algorithms using generative AI agents as behavioral engines for synthetic accounts and establishes GenAI-based agents as a new tool for algorithmic auditing.
Alessandro Morosini, Sarah H. Cen, Andrew Ilyas et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
