Research map / Reinforcement learning / Imitation learning
Self-Adaptive Imitation Learning: Learning Tasks with Delayed Rewards from Sub-optimal Demonstrations
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 |
|---|---|---|---|
| Use SAIL's off-policy exploration instead of curiosity-based exploration | SAIL · HalfCheetah | return: 9043.07 ± 165.97 -> 10660.59±105.53 | Helped |
| Use SAIL's off-policy exploration instead of entropy-based exploration | SAIL · Swimmer | return: 65.04 ± 7.66 -> 309.47±3.0 | Helped |
| Use SAIL's off-policy exploration instead of curiosity-based exploration | SAIL · Hopper | return: 3075.46 ± 15.61 -> 3302.06±14.22 | Helped |
| Use SAIL's off-policy exploration instead of entropy-based exploration | SAIL · Walker | return: 2792.76 ± 830.20 -> 5868.53±108.82 | Helped |
| Sample from teacher demonstrations alongside self-generated data during early learning (ablation) | SAIL · MuJoCo continuous-control tasks | return | Helped |
| Adapt the teacher buffer by replacing it with higher-reward self-generated trajectories (ablation) | SAIL · MuJoCo continuous-control tasks | return | Helped |
| Train the discriminator on on-policy rather than off-policy samples (ablation) | SAIL · MuJoCo continuous-control tasks | return | Hurt |
Closest papers by meaning
- Hybrid Inverse Reinforcement Learning (ICML 2024 · similarity 0.88)
- A Smooth Sea Never Made a Skilled SAILOR: Robust Imitation via Learning to Search (NeurIPS 2025 · similarity 0.88)
- SEABO: A Simple Search-Based Method for Offline Imitation Learning (ICLR 2024 · similarity 0.87)
- RLIF: Interactive Imitation Learning as Reinforcement Learning (ICLR 2024 · similarity 0.87)
- Enhancing Online Reinforcement Learning with Meta-Learned Objective from Offline Data (AAAI 2025 · similarity 0.87)
- Imitation Learning from Observation with Automatic Discount Scheduling (ICLR 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.