A Free-Energy Principle for Representation Learning

Yansong Gao, Pratik Chaudhari
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:3367-3376, 2020.

Abstract

This paper employs a formal connection of machine learning with thermodynamics to characterize the quality of learnt representations for transfer learning. We discuss how information-theoretic functionals such as rate, distortion and classification loss of a model lie on a convex, so-called equilibrium surface. We prescribe dynamical processes to traverse this surface under constraints, e.g., an iso-classification process that trades off rate and distortion to keep the classification loss unchanged. We demonstrate how this process can be used for transferring representations from a source dataset to a target dataset while keeping the classification loss constant. Experimental validation of the theoretical results is provided on standard image-classification datasets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-gao20a, title = {A Free-Energy Principle for Representation Learning}, author = {Gao, Yansong and Chaudhari, Pratik}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {3367--3376}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/gao20a/gao20a.pdf}, url = {https://proceedings.mlr.press/v119/gao20a.html}, abstract = {This paper employs a formal connection of machine learning with thermodynamics to characterize the quality of learnt representations for transfer learning. We discuss how information-theoretic functionals such as rate, distortion and classification loss of a model lie on a convex, so-called equilibrium surface. We prescribe dynamical processes to traverse this surface under constraints, e.g., an iso-classification process that trades off rate and distortion to keep the classification loss unchanged. We demonstrate how this process can be used for transferring representations from a source dataset to a target dataset while keeping the classification loss constant. Experimental validation of the theoretical results is provided on standard image-classification datasets.} }
Endnote
%0 Conference Paper %T A Free-Energy Principle for Representation Learning %A Yansong Gao %A Pratik Chaudhari %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-gao20a %I PMLR %P 3367--3376 %U https://proceedings.mlr.press/v119/gao20a.html %V 119 %X This paper employs a formal connection of machine learning with thermodynamics to characterize the quality of learnt representations for transfer learning. We discuss how information-theoretic functionals such as rate, distortion and classification loss of a model lie on a convex, so-called equilibrium surface. We prescribe dynamical processes to traverse this surface under constraints, e.g., an iso-classification process that trades off rate and distortion to keep the classification loss unchanged. We demonstrate how this process can be used for transferring representations from a source dataset to a target dataset while keeping the classification loss constant. Experimental validation of the theoretical results is provided on standard image-classification datasets.
APA
Gao, Y. & Chaudhari, P.. (2020). A Free-Energy Principle for Representation Learning. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:3367-3376 Available from https://proceedings.mlr.press/v119/gao20a.html.

Related Material