Research map / Reinforcement learning / Policy gradient and actor-critic
Scalable Coordinated Exploration in Concurrent Reinforcement Learning
Maria Dimakopoulou, Ian Osband, Benjamin Van Roy
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use seed sampling instead of episodic PSRL | Concurrent tabular cartpole swing-up · Cartpole swing-up | evaluation reward | Improved |
| Use seed sampling instead of concurrent UCRL | Concurrent tabular cartpole swing-up · Cartpole swing-up | evaluation reward | Improved |
| Use seed sampling instead of Thompson resampling | Concurrent tabular cartpole swing-up · Cartpole swing-up | evaluation reward | Improved |
| Use seed LSVI or seed TD instead of Thompson resampling | Concurrent reinforcement learning · Bipolar chain | mean reward per agent | Improved |
| Use seed LSVI or seed TD instead of concurrent UCRL | Concurrent reinforcement learning · Parallel chains | mean reward per agent | Improved |
| Use seed LSVI or seed TD instead of seed sampling | Concurrent reinforcement learning · Bipolar chain | mean reward per agent | No clear effect |
| Use seed LSVI or seed TD instead of seed sampling | Concurrent reinforcement learning · Parallel chains | mean reward per agent | No clear effect |
| Use seed TD-ensemble instead of DQN with epsilon-greedy exploration | Neural-network concurrent reinforcement learning · High-dimensional cartpole swing-up | average reward per agent | Improved |
| Increase the number of parallel agents | Seed TD-ensemble · High-dimensional cartpole swing-up | average reward per agent | Improved |
| Cap ensemble size at min(K, 30) instead of using a separate model for each agent (ablation) | Seed TD-ensemble · High-dimensional cartpole swing-up | average reward per agent | No clear effect |
Closest papers by meaning
- Coordinated Exploration in Concurrent Reinforcement Learning (ICML 2018 · similarity 0.94)
- Randomized Exploration in Cooperative Multi-Agent Reinforcement Learning (NeurIPS 2024 · similarity 0.88)
- Settling Decentralized Multi-Agent Coordinated Exploration by Novelty Sharing (AAAI 2024 · similarity 0.88)
- Concurrent Reinforcement Learning with Aggregated States via Randomized Least Squares Value Iteration (ICML 2025 · similarity 0.87)
- The Horizon Threshold in Cooperative Multi-Agent Reward-Free Exploration (NeurIPS 2026 · similarity 0.87)
- Random Latent Exploration for Deep Reinforcement Learning (ICML 2024 · similarity 0.87)
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.
