Research map / Reinforcement learning / Imitation learning
Understanding Adversarial Imitation Learning in Small Sample Regime: A Stage-Coupled Analysis
Tian Xu, Ziniu Li, Yang Yu, Zhi-Quan Tom Luo
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 TV-AIL instead of BC | locomotion imitation · Hopper (H=1000, N=1) | scaled imitation gap: 784.97 -> 10.38 | Improved |
| use TV-AIL instead of BC | locomotion imitation · Walker2d (H=1000, N=1) | scaled imitation gap: 1002.13 -> 12.89 | Improved |
| use TV-AIL instead of BC | locomotion imitation · HalfCheetah (H=1000, N=10) | scaled imitation gap: 579.53 -> -79.29 | Improved |
| use TV-AIL instead of BC | RBAS MDP · RBAS MDP (H=1000, N=1) | imitation gap: 998.87 -> 0.71 | Improved |
| increase expert trajectories from 1 to 10 (ablation) | TV-AIL · RBAS MDP (H=1000) | imitation gap: 0.71 -> 0.55 | Improved |
| increase horizon from 100 to 2000 (ablation) | BC · RBAS MDP (N=1) | imitation gap: 98.89 -> 1998.88 | Hurt |
| increase horizon from 100 to 2000 (ablation) | TV-AIL · RBAS MDP (N=1) | imitation gap: 0.69 -> 0.71 | No clear effect |
| increase horizon from 100 to 2000 (ablation) | BC · Hopper (N=1) | scaled imitation gap: 0.80 -> 1950.42 | Hurt |
| increase horizon from 100 to 2000 (ablation) | TV-AIL · Hopper (N=1) | scaled imitation gap: 4.96 -> -9.92 | No clear effect |
| increase horizon (estimation error grows) for TV-AIL (ablation) | TV-AIL · RBAS MDP (N=1) | estimation error vs gap: 189.47 -> 3789.47 estimation error; gap 0.69 -> 0.71 | No clear effect |
| increase expert trajectories from 1 to 10 (ablation) | BC · Walker2d (H=1000) | scaled imitation gap: 1002.13 -> 222.98 | Improved |
Closest papers by meaning
- State-only Imitation with Transition Dynamics Mismatch (ICLR 2020 · similarity 0.92)
- Is Behavior Cloning All You Need? Understanding Horizon in Imitation Learning (NeurIPS 2024 · similarity 0.91)
- Adversarial Imitation Learning With General Function Approximation: Theoretical Analysis and Practical Algorithms (TPAMI 2026 · similarity 0.90)
- Diffusing States and Matching Scores: A New Framework for Imitation Learning (ICLR 2025 · similarity 0.89)
- Adversarial Imitation Learning via Boosting (ICLR 2024 · similarity 0.89)
- Action-Constrained Imitation Learning (ICML 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.
