pith. sign in

arxiv: 1802.07426 · v3 · pith:3RF7L6QEnew · submitted 2018-02-21 · 📊 stat.ML · cs.AI· cs.LG· cs.NE

Generalization in Machine Learning via Analytical Learning Theory

classification 📊 stat.ML cs.AIcs.LGcs.NE
keywords learningtheorydeepstatisticalmachinemethodsproblemproposed
0
0 comments X
read the original abstract

This paper introduces a novel measure-theoretic theory for machine learning that does not require statistical assumptions. Based on this theory, a new regularization method in deep learning is derived and shown to outperform previous methods in CIFAR-10, CIFAR-100, and SVHN. Moreover, the proposed theory provides a theoretical basis for a family of practically successful regularization methods in deep learning. We discuss several consequences of our results on one-shot learning, representation learning, deep learning, and curriculum learning. Unlike statistical learning theory, the proposed learning theory analyzes each problem instance individually via measure theory, rather than a set of problem instances via statistics. As a result, it provides different types of results and insights when compared to statistical learning theory.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.