{"title":"表征学习的自由能原理","authors":"Yansong Gao, P. Chaudhari","doi":"10.1088/2632-2153/ABF984","DOIUrl":null,"url":null,"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 functional 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.","PeriodicalId":18148,"journal":{"name":"Mach. Learn. Sci. Technol.","volume":"36 1","pages":"45004"},"PeriodicalIF":0.0000,"publicationDate":"2020-02-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":"{\"title\":\"A Free-Energy Principle for Representation Learning\",\"authors\":\"Yansong Gao, P. Chaudhari\",\"doi\":\"10.1088/2632-2153/ABF984\",\"DOIUrl\":null,\"url\":null,\"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 functional 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.\",\"PeriodicalId\":18148,\"journal\":{\"name\":\"Mach. Learn. Sci. Technol.\",\"volume\":\"36 1\",\"pages\":\"45004\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2020-02-27\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"7\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Mach. Learn. Sci. Technol.\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1088/2632-2153/ABF984\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Mach. Learn. Sci. Technol.","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1088/2632-2153/ABF984","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
A Free-Energy Principle for Representation Learning
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 functional 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.