WRSWorld Research StateContribute
WORLD RESEARCH STATE

MOReL: unknown-region pessimistic MDP construction for offline learning

Contribution grade

Offline model-based reinforcement learning · Human-designed research method or resource

Reference score115,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Partitioned state-action space using an unknown-state-action detector, redirected unknown transitions to an absorbing penalty state and planned in the resulting pessimistic MDP.

Scientific scope

Partitioned state-action space using an unknown-state-action detector, redirected unknown transitions to an absorbing penalty state and planned in the resulting pessimistic MDP. Original theorem includes initial-distribution and known-region total-variation error terms; the abstract's unqualified lower-bound wording is not exported. Detector guarantees, planning accuracy and coverage/hitting-time terms matter. Practical detectors are not automatically certified error oracles.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Original theorem includes initial-distribution and known-region total-variation error terms; the abstract's unqualified lower-bound wording is not exported. Detector guarantees, planning accuracy and coverage/hitting-time terms matter. Practical detectors are not automatically certified error oracles.

Contributors

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