The paper's claimed O(SA log A) sample-complexity bound for RL in metric state spaces is not supported due to a lower bound being misused as an upper bound in the proof.
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
-
Improving Reinforcement Learning Sample-Efficiency using Local Approximation
The paper's claimed O(SA log A) sample-complexity bound for RL in metric state spaces is not supported due to a lower bound being misused as an upper bound in the proof.