Research map / Reinforcement learning / Offline RL
Reinforcement Learning with Sparse Rewards using Guidance from Offline Demonstration
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 |
|---|---|---|---|
| add offline demonstration guidance with a decaying trust region to TRPO (LOGO) | TRPO in sparse-reward MuJoCo tasks · MuJoCo sparse-reward environments | reward | Helped |
| use TRPO without behavior-data guidance | TRPO · MuJoCo sparse-reward environments | reward | Hurt |
| imitate the behavior policy with GAIL instead of using online policy guidance | GAIL · MuJoCo sparse-reward environments | reward | Hurt |
| warm-start TRPO with behavior cloning, without ongoing online guidance | BC-TRPO · MuJoCo sparse-reward environments | reward | Mixed |
| use POfD to bootstrap learning from behavior data, retaining its influence throughout training | POfD · MuJoCo sparse-reward environments | reward | Hurt |
| use LOGO rather than DAPG | LOGO and DAPG · MuJoCo sparse-reward environments | reward | Mixed |
| use LOGO with incomplete offline state observations | LOGO · MuJoCo sparse-reward environments with incomplete offline observation | reward | Mixed |
| use offline behavior-policy guidance with LOGO | LOGO · TurtleBot waypoint tracking | waypoint-tracking performance | Helped |
| use sparse-reward TRPO without behavior-data guidance | TRPO · TurtleBot waypoint tracking | waypoint attainment | Hurt |
| guide LOGO with waypoint behavior data that lacks obstacle information, while supplying the full state during learning | LOGO · TurtleBot obstacle avoidance | obstacle-avoidance performance | Helped |
| use sparse-reward TRPO without behavior-data guidance | TRPO · TurtleBot obstacle avoidance | collision avoidance | Hurt |
Closest papers by meaning
- DISCOVER: Automated Curricula for Sparse-Reward Reinforcement Learning (NeurIPS 2025 · similarity 0.89)
- Offline RL with Observation Histories: Analyzing and Improving Sample Complexity (ICLR 2024 · similarity 0.89)
- Learning Process Rewards via Visitation Matching for Efficient RL (NeurIPS 2026 · similarity 0.89)
- A Clean Slate for Offline Reinforcement Learning (NeurIPS 2025 · similarity 0.89)
- Iteratively Refined Behavior Regularization for Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data (AAAI 2025 · similarity 0.89)
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.