High-arity PAC learning via exchangeability

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Coregliano, Leonardo N., Malliaris, Maryanthe
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914950341984256
author Coregliano, Leonardo N.
Malliaris, Maryanthe
author_facet Coregliano, Leonardo N.
Malliaris, Maryanthe
contents We develop a theory of high-arity PAC learning, which is statistical learning in the presence of "structured correlation". In this theory, hypotheses are either graphs, hypergraphs or, more generally, structures in finite relational languages, and i.i.d. sampling is replaced by sampling an induced substructure, producing an exchangeable distribution. Our main theorems establish a high-arity (agnostic) version of the fundamental theorem of statistical learning.
format Preprint
id arxiv_https___arxiv_org_abs_2402_14294
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle High-arity PAC learning via exchangeability
Coregliano, Leonardo N.
Malliaris, Maryanthe
Machine Learning
Logic
Statistics Theory
Primary: 68Q32. Secondary: 60F05, 60F15, 03C99
We develop a theory of high-arity PAC learning, which is statistical learning in the presence of "structured correlation". In this theory, hypotheses are either graphs, hypergraphs or, more generally, structures in finite relational languages, and i.i.d. sampling is replaced by sampling an induced substructure, producing an exchangeable distribution. Our main theorems establish a high-arity (agnostic) version of the fundamental theorem of statistical learning.
title High-arity PAC learning via exchangeability
topic Machine Learning
Logic
Statistics Theory
Primary: 68Q32. Secondary: 60F05, 60F15, 03C99
url https://arxiv.org/abs/2402.14294