Research map / Reinforcement learning / Imitation learning
Learning Noise-Induced Reward Functions for Surpassing Demonstrations in Imitation Learning
Liangyu Huo, Zulin Wang, Mai Xu
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 weighted preference learning with a nonlinear weight based on noise-level differences | LERP · HalfCheetah, Hopper, Walker2d, Humanoid, Enduro, Kung Fu Master, Pong, Seaquest, Space Invaders, Ms Pacman | average return | Improved |
| Use weighted preference learning with a nonlinear weight based on noise-level differences | LERP · Ant | average return: 738.9 -> 612.7 | Hurt |
| Use weighted preference learning with a nonlinear weight based on noise-level differences | LERP · Q*bert | average return: 27967.5 -> 9622.5 | Hurt |
| Replace the nonlinear noise weight with a linear weight (ablation) | LERP · HalfCheetah unseen demonstrations | return correlation: 0.830 -> 0.899 | Improved |
| Replace the nonlinear noise weight with a linear weight (ablation) | LERP · Hopper unseen demonstrations | return correlation: -0.830 -> -0.624 | Improved |
| Replace the nonlinear noise weight with a linear weight (ablation) | LERP · Ant unseen demonstrations | return correlation: 0.913 -> 0.963 | Improved |
| Replace the nonlinear noise weight with a linear weight (ablation) | LERP · Walker2d unseen demonstrations | return correlation: 0.421 -> -0.122 | Hurt |
| Replace the nonlinear noise weight with a linear weight (ablation) | LERP · Humanoid unseen demonstrations | return correlation: 0.608 -> 0.964 | Improved |
| Use the nonlinear noise weight rather than the linear-weight variant (ablation) | LERP reward model · MuJoCo preference pairs | preference accuracy: 76.3% -> 82.0% | Improved |
| Use the nonlinear noise weight rather than the linear-weight variant (ablation) | LERP reward model · MuJoCo preference pairs | preference accuracy: 89.9% -> 92.7% | Improved |
Closest papers by meaning
- Limited Preference Aided Imitation Learning from Imperfect Demonstrations (ICML 2024 · similarity 0.89)
- Uncertainty-Based Smooth Policy Regularisation for Reinforcement Learning with Few Demonstrations (NeurIPS 2025 · similarity 0.88)
- Imitation Learning from Purified Demonstrations (ICML 2024 · similarity 0.87)
- Markov Balance Satisfaction Improves Performance in Strictly Batch Offline Imitation Learning (AAAI 2025 · similarity 0.87)
- Imitation Learning from Observation with Automatic Discount Scheduling (ICLR 2024 · similarity 0.87)
- Improving Reward Model Generalization from Adversarial Process Enhanced Preferences (ICML 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.
