AC0 admits quasipolynomial-time learning under Gibbs measures with efficient local samplers, covering hard-core and Ising models on arbitrary bounded-degree graphs near sampling thresholds.
Toward Derandomizing
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Learning $\mathsf{AC}^0$ under Locally Sampleable Graphical Models
AC0 admits quasipolynomial-time learning under Gibbs measures with efficient local samplers, covering hard-core and Ising models on arbitrary bounded-degree graphs near sampling thresholds.