Research map / Reinforcement learning / Policy gradient and actor-critic
Increasing the Action Gap: New Operators for Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Replace Bellman updates with advantage learning updates | DQN · Original DQN, across games | median score improvement over DQN: 8.4% | Helped |
| Replace Bellman updates with persistent advantage learning updates | DQN · Original DQN, across games | median score improvement over DQN: 9.1% | Helped |
| Replace Bellman updates with advantage learning updates | DQN · Original DQN, across games | average score improvement over DQN: 27.0% | Helped |
| Replace Bellman updates with persistent advantage learning updates | DQN · Original DQN, across games | average score improvement over DQN: 32.5% | Helped |
| Replace Bellman updates with advantage learning updates | DQN · Stochastic Minimal, 60 Atari games | games statistically better than DQN: 37 | Helped |
| Replace Bellman updates with persistent advantage learning updates | DQN · Stochastic Minimal, 60 Atari games | games statistically better than DQN: 35 | Helped |
| Use persistent advantage learning instead of advantage learning | DQN · Original DQN, across Atari games | score | Mixed |
| Use consistent Q-value interpolation instead of ordinary Q-value interpolation | Bicycle value iteration · Bicycle domain, coarse grid | goal-reaching performance | Helped |
| Add the persistent-action term to advantage learning | DQN · Atari 2600 games | score: 16 games; 6 games | Mixed |
Closest papers by meaning
- Gradual Transition from Bellman Optimality Operator to Bellman Operator in Online Reinforcement Learning (ICML 2025 · similarity 0.89)
- Epistemic Bellman Operators (AAAI 2025 · similarity 0.88)
- Taming "data-hungry" reinforcement learning? Stability in continuous state-action spaces (NeurIPS 2024 · similarity 0.87)
- Inverting the Bellman Equation: From $Q$-Values to World Models (NeurIPS 2026 · similarity 0.87)
- Natural Policy Gradient as Doubly Smoothed Policy Iteration: A Bellman-Operator Framework (NeurIPS 2026 · similarity 0.86)
- Parameterized Projected Bellman Operator (AAAI 2024 · similarity 0.86)
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.