Value mirror descent integrates mirror descent into value iteration for discounted MDPs, delivering near-optimal sample complexity of order |S||A|(1-γ)^{-3}ε^{-2} for general convex regularizers and bounded Bregman divergence between generated and optimal policies.
Policy mirror descent inherently explores action space.SIAM Journal on Optimization, 35(1):116–156, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Value Mirror Descent for Reinforcement Learning
Value mirror descent integrates mirror descent into value iteration for discounted MDPs, delivering near-optimal sample complexity of order |S||A|(1-γ)^{-3}ε^{-2} for general convex regularizers and bounded Bregman divergence between generated and optimal policies.