BWhat changed
Exposes an actor's unperturbed greedy action to teammates during centralized training so recurrent agents can decode action information without epsilon-greedy exploration blur; greedy execution removes that side channel at test time.
Scientific scope
Exposes an actor's unperturbed greedy action to teammates during centralized training so recurrent agents can decode action information without epsilon-greedy exploration blur; greedy execution removes that side channel at test time. Implicit recurrent beliefs are not exact Bayesian posteriors. Directly observed actions are the straightforward case; indirectly observed actions require an additional inverse-model extension. Inherited DQN/VDN/distributed training is not newly invented, and reading SAD does not resolve the separately held R2D2 original-source gap.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Implicit recurrent beliefs are not exact Bayesian posteriors. Directly observed actions are the straightforward case; indirectly observed actions require an additional inverse-model extension. Inherited DQN/VDN/distributed training is not newly invented, and reading SAD does not resolve the separately held R2D2 original-source gap.
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