Research map / Reinforcement learning / Visual and representation RL
Learning to Execute: Efficient Learning of Universal Plan-Conditioned Policies in Robotics
Ingmar Schubert, Danny Driess, Ozgur S. Oguz, Marc Toussaint
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 |
|---|---|---|---|
| add plan-conditioned policy with plan-based reward shaping (L2E) versus pure RL (HER), pure planning, plan+inverse model, planned subgoals + RL | L2E · Basic pushing | success rate | Improved |
| add plan-conditioned policy (L2E) versus HER | L2E · Obstacle pushing | success rate | Improved |
| use plan-conditioned L2E instead of goal-conditioned HER in a quickly changing environment | L2E · Moving obstacles pushing | success rate | Improved |
| use biased replay strategy instead of uniform replay (ablation) | L2E · Basic pushing | sample efficiency | Improved |
| use any plan encoding versus no encoding (ablation) | L2E S^bias_{10,1000} · Basic pushing | success rate | Improved |
| learned VAE encoding instead of analytical 4D encoding (ablation) | L2E · Basic pushing | success rate | Hurt |
| reduce plan density from 50 states to 12 states (ablation) | L2E · Basic pushing | success rate | No clear effect |
| reduce plan density to 3 states (ablation) | L2E · Basic pushing | success rate | Hurt |
| HER replay strategy Future 1 versus Episode 1, Episode 5 (ablation) | HER · Basic pushing | learning speed | Improved |
| larger plan space (6D latent in obstacle task versus 4D in basic task) | L2E · Obstacle pushing | learning speed | Hurt |
Closest papers by meaning
- Efficient Robotic Policy Learning via Latent Space Backward Planning (ICML 2025 · similarity 0.88)
- Plan-Seq-Learn: Language Model Guided RL for Solving Long Horizon Robotics Tasks (ICLR 2024 · similarity 0.88)
- DUO: Diverse, Uncertain, On-Policy Query Generation and Selection for Reinforcement Learning from Human Feedback (AAAI 2025 · similarity 0.87)
- Multi-Stage Manipulation with Demonstration-Augmented Reward, Policy, and World Model Learning (ICML 2025 · similarity 0.87)
- Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes (AAAI 2025 · similarity 0.87)
- Oracle-Efficient Reinforcement Learning for Max Value Ensembles (NeurIPS 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.
