Gorila: distributed actors, learners and parameters for deep Q-learning

What changed

Distributed experience generation, replay, gradient computation and parameter storage across actor/learner processes to scale deep Q-network training.

Scientific scope

Distributed experience generation, replay, gradient computation and parameter storage across actor/learner processes to scale deep Q-network training. Inherits DQN, replay and parameter-server/distributed optimization designs. This original Gorila architecture is distinct from later IMPALA/Ape-X but does not establish optimal scaling or communication-free learning; wall-clock comparisons use substantial parallel resources.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits DQN, replay and parameter-server/distributed optimization designs. This original Gorila architecture is distinct from later IMPALA/Ape-X but does not establish optimal scaling or communication-free learning; wall-clock comparisons use substantial parallel resources.

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.

• Massively Parallel Methods for Deep Reinforcement Learning

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