rl-convergence-prover

1

chuongdlb

Proves or disproves convergence of RL algorithms using contraction mapping, Dvoretzky, Extended Dvoretzky, and Robbins-Monro theorems.

rl-theory-analyzer

1

chuongdlb

Analyzes RL algorithms mathematically — derives update rules, proves convergence, identifies failure modes, and classifies algorithms along standard taxonomic dimensions.