AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

Accelerating Q-learning through Efficient Value-Sharing across Actions

Prabhat Nagarajan, Brett Daley, Martha White, Marlos C. Machado

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Temporal-difference and discounting theory · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add the mean-expansion layer (k=n) to DQNDQN · Atari-57IQM human-normalized score: matches DQN's best performance at just under 20M timesteps, within 40% of training timeImproved
add the mean-expansion layer to IQNIQN · Atari-57IQM human-normalized score: reaches IQN's highest score at 33.75M timestepsImproved
add the mean-expansion layer to tabular Q-learning (IBQ)Q-learning · 5x5 stochastic gridworldcompleted episodes: over 20% more episodes within 1k timestepsImproved
use mean-scaling k=n instead of k=1 (ablation)IB-DQN · Atari-57IQM human-normalized scoreImproved
increase k beyond a moderate value with fixed step size (ablation)IB-DQN · Atari and LunarLander-v3AUC of returnHurt
add Dueling DQN architecture to DQNDQN · Atari-57IQM human-normalized scoreNo clear effect
add the mean-expansion layer to DQN with RMSprop and Huber lossDQN · Atari (55 games)IQM human-normalized scoreImproved
add the mean-expansion layer to DQNDQN · Atari-57value overestimation: reduced in all gamesImproved
add the mean-expansion layer to DQNDQN · Atari-57relative action gapImproved
add the mean-expansion layer (k=n) to DQNDQN · Atari Defendermean score: 9260.5 -> 22271.0Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy