AWhat changed
Combined bounded importance-ratio traces with corrected multi-step Bellman returns, retaining long near-on-policy returns while controlling off-policy trace growth.
Scientific scope
Combined bounded importance-ratio traces with corrected multi-step Bellman returns, retaining long near-on-policy returns while controlling off-policy trace growth. Safe refers to off-policy value-operator convergence, not real-world deployment safety. Expected fixed-policy contraction uses nonnegative trace coefficients bounded by target/behavior likelihood ratios; control and stochastic convergence require additional greedy-policy, initialization, coverage and learning conditions. No guarantee for arbitrary neural approximators or absent exploration.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Safe refers to off-policy value-operator convergence, not real-world deployment safety. Expected fixed-policy contraction uses nonnegative trace coefficients bounded by target/behavior likelihood ratios; control and stochastic convergence require additional greedy-policy, initialization, coverage and learning conditions. No guarantee for arbitrary neural approximators or absent exploration.
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.
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