# High-arity PAC learning via exchangeability

arXiv: https://arxiv.org/abs/2402.14294 · HTML: https://mlsanity.com/paper/2402.14294/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 1 of its 26 references that are in the map.
- First posted on arXiv 22 Feb 2024.

## Builds on (most cited first, 1 of 1)

- [Learning from non-irreducible Markov chains](https://mlsanity.com/paper/2110.04338/) (2021)

## Built on by (most cited first, 4 of 4)

- [Uniform Laws of Large Numbers in Product Spaces](https://mlsanity.com/paper/2603.24493/) (2026)
- [Higher-arity PAC learning, VC dimension and packing lemma](https://mlsanity.com/paper/2510.02420/) (2025)
- [On statistical learning of graphs](https://mlsanity.com/paper/2507.13054/) (2025)
- [Remarks on a recent preprint of Chernikov and Towsner](https://mlsanity.com/paper/2510.19665/) (2025)

## Abstract (arXiv)

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.