MODEL THEORY AND MACHINE LEARNING

DOI10.1017/BSL.2018.71zbMATH Open1506.03095arXiv1801.06566OpenAlexW2785170103MaRDI QIDQ5240809FDOQ5240809

Authors: Hunter Chase, James Freitag 

Publication date: 29 October 2019

Published in: The Bulletin of Symbolic Logic (Search for Journal in Brave)

Abstract: About 25 years ago, it came to light that a single combinatorial property determines both an important dividing line in model theory (NIP) and machine learning (PAC-learnability). The following years saw a fruitful exchange of ideas between PAC learning and the model theory of NIP structures. In this article, we point out a new and similar connection between model theory and machine learning, this time developing a correspondence between emph{stability} and learnability in various settings of emph{online learning.} In particular, this gives many new examples of mathematically interesting classes which are learnable in the online setting.

Full work available at URL: https://arxiv.org/abs/1801.06566

Recommendations

• scientific article; zbMATH DE number 939032
• Model theory
• scientific article; zbMATH DE number 5296405
• scientific article; zbMATH DE number 53151
• Model theory
• scientific article; zbMATH DE number 798660
• scientific article; zbMATH DE number 44997
• Mathematical theories of machine learning: theory and applications
• scientific article; zbMATH DE number 6528605
• Model learning as a satisfiability modulo theories problem

zbMATH Keywords

machine learningmodel theory

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Classification theory, stability, and related concepts in model theory (03C45) Applications of model theory (03C98)

Cites Work

• The weighted majority algorithm
• Title not available (Why is that?)
• Model theory for metric structures
• Classification theory and the number of non-isomorphic models
• Vapnik-Chervonenkis Classes of Definable Sets
• Externally definable sets and dependent pairs
• Notes on the stability of separably closed fields
• Title not available (Why is that?)
• Elimination of quantifiers for modules
• Title not available (Why is that?)
• Compression schemes, stable definable families, and o-minimal structures
• Model theory and complex geometry

Cited In (10)

• Regular partitions of gentle graphs
• Enhancing deep learning algorithm accuracy and stability using multicriteria optimization: an application to distributed learning with MNIST digits
• Title not available (Why is that?)
• THICKET DENSITY
• Machine Learning and Invariant Theory
• Model theory and agnostic online learning via excellent sets
• Machine learning and the continuum hypothesis
• The amalgamation property for automorphisms of ordered abelian groups
• Title not available (Why is that?)
• MODEL THEORY AND COMBINATORICS OF BANNED SEQUENCES

This page was built for publication: MODEL THEORY AND MACHINE LEARNING

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5240809)