<ul><li>Authors introduced high-arity PAC learning theory for graphs, hypergraphs, and relational structures.</li><li>They proved a high-arity analogue of the Fundamental Theorem of Statistical Learning regarding Vapnik-Chervonenkis-Natarajan (VCN$_k$) $k$-dimension.</li><li>The study completed the characterization by showing non-partite non-agnostic high-arity PAC learnability leads to a version of Haussler packing property.</li><li>Proofs established classic PAC learnability implies classic Haussler packing property and finiteness of VCN$_k$-dimension, extending to high-arity scenarios.</li></ul>

A packing lemma for VCN${}_k$-dimension and learning high-dimensional data

Discover more