Research map / Reinforcement learning / Visual and representation RL
TORA: Train Once, Realign Anytime for Offline Multi-Objective Reinforcement Learning
Weichen Li, Waleed Mustafa, Marcio Monteiro, Puyu Wang, Marius Kloft, Sophie Fellenz
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 TORA composed diffusion instead of MODT | D4MORL expert Ant · D4MORL Ant | HV: 5.52 -> 6.21 | Improved |
| use TORA instead of CQL | D4MORL expert Walker2d · D4MORL Walker2d | HV: 0.82 -> 5.16 | Improved |
| use TORA instead of MORvS | D4MORL expert Swimmer · D4MORL Swimmer | HV: 3.19 -> 3.17 | No clear effect |
| use TORA instead of MODT | D4MORL expert HalfCheetah · D4MORL HalfCheetah | HV: 5.59 -> 5.71 | No clear effect |
| use TORA (no training-time preferences) instead of MODULI | D4MORL expert · D4MORL | HV | No clear effect |
| use TORA instead of OASIS | DroneRun · Bullet-Safety-Gym DroneRun | normalized reward: 0.13 -> 0.27 | Improved |
| use TORA instead of OASIS | DroneRun · Bullet-Safety-Gym DroneRun | normalized cost: 0.79 -> 0.00 | Improved |
| use TORA instead of OASIS | DroneCircle · Bullet-Safety-Gym DroneCircle | normalized reward: 0.60 -> 0.16 | Hurt |
| use TORA instead of OASIS | CarCircle · Bullet-Safety-Gym CarCircle | normalized reward: 0.76 -> 0.42 | Hurt |
| use TORA instead of CDT | DroneRun · Bullet-Safety-Gym DroneRun | normalized cost: 7.56 -> 0.00 | Improved |
Closest papers by meaning
- MODULI: Unlocking Preference Generalization via Diffusion Models for Offline Multi-Objective Reinforcement Learning (ICML 2025 · similarity 0.89)
- An Offline Adaptation Framework for Constrained Multi-Objective Reinforcement Learning (NeurIPS 2024 · similarity 0.88)
- Preference Conditioned Multi-Objective Reinforcement Learning: Decomposed, Diversity-Driven Policy Optimization (NeurIPS 2026 · similarity 0.88)
- On Generalization Across Environments In Multi-Objective Reinforcement Learning (ICLR 2025 · similarity 0.87)
- An Analytical Study of Utility Functions in Multi-Objective Reinforcement Learning (NeurIPS 2024 · similarity 0.86)
- Pareto Set Learning for Multi-Objective Reinforcement Learning (AAAI 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.
