AWhat changed
Combines recurrent Q-learning and prioritized distributed sequence replay, storing recurrent states and using burn-in to reduce training-state mismatch and representational staleness.
Scientific scope
Combines recurrent Q-learning and prioritized distributed sequence replay, storing recurrent states and using burn-in to reduce training-state mismatch and representational staleness. The record concerns the published recurrent-replay method, rather than a mixture of draft-specific performance figures. It builds on Ape-X, DQN, recurrent networks and prioritized replay. Burn-in and stored states mitigate staleness empirically; they do not guarantee convergence or recovery of an arbitrary complete history.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. The record concerns the published recurrent-replay method, rather than a mixture of draft-specific performance figures. It builds on Ape-X, DQN, recurrent networks and prioritized replay. Burn-in and stored states mitigate staleness empirically; they do not guarantee convergence or recovery of an arbitrary complete history.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- Recurrent Experience Replay in Distributed Reinforcement Learning —18-page review draft
- Recurrent Experience Replay in Distributed Reinforcement Learning — author publication record
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction