RESPONSIBILITY IN AI SYSTEMS & EXPERIENCES UNIVERSITY of WASHINGTON
Neural NetworksAccountabilityHuman-centered AI

Effective Backdoor Mitigation Depends on the Pre-training Objective

arXiv e-prints ·
Sahil Verma, Gantavya Bhatt, Avi Schwarzschild, Soumye Singhal, Arnav Mohanty Das, Chirag Shah, John P. Dickerson, Jeff Bilmes
View on DOI ↗

Cite as

Sahil Verma, Gantavya Bhatt, Avi Schwarzschild, Soumye Singhal, Arnav Mohanty Das, Chirag Shah, John P. Dickerson, Jeff Bilmes. (2023). Effective Backdoor Mitigation Depends on the Pre-training Objective. arXiv e-prints. https://doi.org/10.48550/arxiv.2311.14948