2 papers
cs.LG2026
Influence Dynamics and Stagewise Data Attribution
Jin Hwa Lee, Matthew Smith, Maxwell Adam +1
Current training data attribution (TDA) methods treat the influence one sample has on another as static, but neural networks learn in distinct stages that exhibit changing patterns…
cs.LG2026
Bayesian Influence Functions for Hessian-Free Data Attribution
Philipp Alexander Kreer, Wilson Wu, Maxwell Adam +2
Classical influence functions face significant challenges when applied to deep neural networks, primarily due to non-invertible Hessians and high-dimensional parameter spaces. We p…