Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Domain-Robust Visual Imitation Learning with Mutual Information Constraints
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 the prior data constraint (ablation) | DisentanGAIL · Inverted Pendulum and Reacher realms with domain mismatches | normalized cumulative reward: Reacher: 0.586\pm 0.143 -> 0.885\pm 0.064; Inverted Pendulum: 0.794\pm 0.234 -> 0.894\pm 0.231 | Helped |
| add the prior data constraint (ablation) | DisentanGAIL · Half-Cheetah | normalized cumulative reward: 0.660\pm 0.231 -> 0.712\pm 0.036 | No clear effect |
| replace mutual information constraints with domain confusion loss (ablation) | DisentanGAIL · Inverted Pendulum and Reacher realms | normalized cumulative reward: Reacher with embodiment and appearance mismatch: 0.860\pm 0.081 -> 0.523\pm 0.177 | Hurt |
| remove latent representation regularization (ablation) | DisentanGAIL · Inverted Pendulum and Reacher with domain mismatches | normalized cumulative reward: Reacher with embodiment and appearance mismatch: 0.860\pm 0.081 -> 0.186\pm 0.136 | Hurt |
| tighten the expert demonstrations constraint from I_max=0.99 to I_max=0.75 (ablation) | DisentanGAIL · Inverted Pendulum and Reacher realms | normalized cumulative reward: Reacher, embodiment: 0.941\pm 0.045 -> 0.841\pm 0.113; Reacher, appearance: 0.885\pm 0.064 -> 0.903\pm 0.064 | Mixed |
Closest papers by meaning
- Learn what matters: cross-domain imitation learning with task-relevant embeddings (NeurIPS 2022 · similarity 0.90)
- Imitation Learning via Kernel Mean Embedding (AAAI 2018 · similarity 0.90)
- Latent Wasserstein Adversarial Imitation Learning (ICLR 2026 · similarity 0.90)
- InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations (NeurIPS 2017 · similarity 0.89)
- Unlabeled Imperfect Demonstrations in Adversarial Imitation Learning (AAAI 2023 · similarity 0.89)
- Diffusing States and Matching Scores: A New Framework for Imitation Learning (ICLR 2025 · similarity 0.88)
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.