Information-theoretic generalization bounds for black-box learning algorithmsDownload PDF

May 21, 2021 (edited Dec 21, 2021)NeurIPS 2021 PosterReaders: Everyone
  • Keywords: generalization, information theory, deep learning theory, stability, memorization
  • TL;DR: We derive information-theoretic generalization bounds for supervised learning algorithms based on the information contained in predictions rather that in the output of the training algorithm.
  • Abstract: We derive information-theoretic generalization bounds for supervised learning algorithms based on the information contained in predictions rather than in the output of the training algorithm. These bounds improve over the existing information-theoretic bounds, are applicable to a wider range of algorithms, and solve two key challenges: (a) they give meaningful results for deterministic algorithms and (b) they are significantly easier to estimate. We show experimentally that the proposed bounds closely follow the generalization gap in practical scenarios for deep learning.
  • Supplementary Material: pdf
  • Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.
  • Code: https://github.com/hrayrhar/f-CMI
14 Replies

Loading