Research map / Reinforcement learning / Imitation learning
Generalizable Imitation Learning from Observation via Inferring Goal Proximity
Youngwoon Lee, Andrew Szot, Shao-Hua Sun, Joseph J Lim
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 uncertainty penalty to GAIfO-s (ablation) | GAIfO-s · average over four tasks | success rate: 18.4% | Improved |
| add goal proximity with difference reward (no uncertainty) (ablation) | GAIfO-s · average over four tasks | success rate: 66.7% | Improved |
| add goal proximity with absolute reward (no uncertainty) (ablation) | GAIfO-s · average over four tasks | success rate: 25.8% | Improved |
| add uncertainty penalty to proximity with difference reward (ablation) | Prox+Diff · average over four tasks | success rate: 66.7% -> 68.7% | Improved |
| add uncertainty penalty to proximity with absolute reward (ablation) | Prox+Abs · average over four tasks | success rate: 25.8% -> 26.4% | Improved |
| use ensemble of discriminators in baseline (ablation) | GAIfO-s · average over four tasks | success rate: 19.6% | Improved |
| apply spectral normalization to discriminator (ablation) | GAIfO-s · FETCH PICK 1.75x noise | success rate: 1.43% -> 14.56% | Improved |
| replace spectral-normalized baseline with goal proximity method | GAIfO-s · FETCH PICK 1.75x noise | success rate: 14.56% -> 75.45% | Improved |
| use absolute instead of difference-based proximity reward (ablation) | Ours · ANT REACH and HAND ROTATE | success rate | Mixed |
| vary uncertainty penalty coefficient away from 0.001 (ablation) | Ours · FETCH PICK 1.75x | success rate | Hurt |
| remove online proximity function update (ablation) | Ours · FETCH PICK 1.75x | success rate | Hurt |
| remove offline pre-training of proximity function (ablation) | Ours (No Offline) · FETCH PICK 1.75x | training speed | Hurt |
Closest papers by meaning
- Learning from Demonstrations via Capability-Aware Goal Sampling (NeurIPS 2025 · similarity 0.90)
- Imitation Learning from Observation with Automatic Discount Scheduling (ICLR 2024 · similarity 0.90)
- GO-DICE: Goal-Conditioned Option-Aware Offline Imitation Learning via Stationary Distribution Correction Estimation (AAAI 2024 · similarity 0.90)
- Zero-Shot Offline Imitation Learning via Optimal Transport (ICML 2025 · similarity 0.89)
- Expert Proximity as Surrogate Rewards for Single Demonstration Imitation Learning (ICML 2024 · similarity 0.89)
- OSVI-WM: One-Shot Visual Imitation for Unseen Tasks using World-Model-Guided Trajectory Generation (NeurIPS 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.
