Research map / Reinforcement learning / Visual and representation RL
Learning General World Models in a Handful of Reward-Free Deployments
Yingchen Xu, Jack Parker-Holder, Aldo Pacchiano, Philip J. Ball, Oleh Rybkin, Stephen J. Roberts, Tim Rocktäschel, Edward Grefenstette
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 |
|---|---|---|---|
| Replace single-policy Plan2Explore with a population of diverse explorers (PP2E) | Plan2Explore · Crafter, 20 deployments of 50k steps | Crafter Score | Improved |
| Add diversity objective (PopDiv) on top of population of information-gain explorers | PP2E · MiniGrid MultiRoom | rewarding episodes | Improved |
| Add diversity objective on top of PP2E | PP2E · MiniGrid FourRooms | rewarding episodes | No clear effect |
| Use CASCADE instead of random exploration | World model exploration · MiniGrid zero-shot | success rate: 0% -> almost 100% | Improved |
| Use CASCADE instead of baselines | DreamerV2 exploration · Atari Montezuma's Revenge, Frostbite, Hero, Freeway | zero-shot return | Improved |
| Use CASCADE instead of PP2E | DMC walker exploration · DMC walker stand/walk/run/flip, 15 deployments | IQM normalized score | Improved |
| Use single-agent P2E instead of populations | DMC walker · DMC walker | zero-shot performance | Hurt |
| Use medium initial dataset instead of random or expert | CASCADE · DMC walker | zero-shot performance | Improved |
| Random exploration versus information-gain baselines | Random exploration · DMC behaviors vs deep exploration worlds | downstream performance | Mixed |
Closest papers by meaning
- Efficient Exploration and Discriminative World Model Learning with an Object-Centric Abstraction (ICLR 2025 · similarity 0.87)
- Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models (NeurIPS 2026 · similarity 0.87)
- Decoupling Exploration and Policy Optimization: Uncertainty Guided Tree Search for Hard Exploration (NeurIPS 2026 · similarity 0.87)
- Towards General-Purpose Model-Free Reinforcement Learning (ICLR 2025 · similarity 0.87)
- Learning World Models for Unconstrained Goal Navigation (NeurIPS 2024 · similarity 0.87)
- Learning from Reward-Free Offline Data: A Case for Planning with Latent Dynamics Models (NeurIPS 2025 · 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.
