A quantum natural policy gradient algorithm with deterministic truncated estimators achieves tilde O(epsilon^{-1.5}) sample complexity for infinite-horizon model-free RL, improving on the classical tilde O(epsilon^{-2}) rate.
O., Ghosh, A., and Aggarwal, V
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
quant-ph 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Accelerating Quantum Reinforcement Learning with a Quantum Natural Policy Gradient Based Approach
A quantum natural policy gradient algorithm with deterministic truncated estimators achieves tilde O(epsilon^{-1.5}) sample complexity for infinite-horizon model-free RL, improving on the classical tilde O(epsilon^{-2}) rate.