Under variational gradient dominance, the paper derives state-space-independent sample complexity bounds for SDPO, CPI, DA-CPI, and PMD in agnostic policy learning.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Convergence and Sample Complexity of First-Order Methods for Agnostic Reinforcement Learning
Under variational gradient dominance, the paper derives state-space-independent sample complexity bounds for SDPO, CPI, DA-CPI, and PMD in agnostic policy learning.