The Memory Perturbation Equation: Understanding Model's Sensitivity to Data
Abstract
Understanding model's sensitivity to its training data is crucial but can also be challenging and costly, especially during training. To simplify such issues, we present the Memory-Perturbation Equation (MPE) which relates model's sensitivity to perturbation in its training data. Derived using Bayesian principles, the MPE unifies existing sensitivity measures, generalizes them to a wide-variety of models and algorithms, and unravels useful properties regarding sensitivities. Our empirical results show that sensitivity estimates obtained during training can be used to faithfully predict generalization on unseen test data. The proposed equation is expected to be useful for future research on robust and adaptive learning.
Cite
@article{arxiv.2310.19273,
title = {The Memory Perturbation Equation: Understanding Model's Sensitivity to Data},
author = {Peter Nickl and Lu Xu and Dharmesh Tailor and Thomas Möllenhoff and Mohammad Emtiyaz Khan},
journal= {arXiv preprint arXiv:2310.19273},
year = {2024}
}
Comments
37th Conference on Neural Information Processing Systems (NeurIPS 2023)