Research map / Vision-language and multimodal / Robot action and navigation
Multi-Modal Imitation Learning from Unstructured Demonstrations using Generative Adversarial Nets
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 latent intention cost to the imitation objective | Imitation GAN · Reacher, 2 targets | target modes learned | Helped |
| Add the latent intention cost to the imitation objective | Imitation GAN · Walker-2D | expert behaviors imitated | Helped |
| Add the latent intention cost to the imitation objective | Imitation GAN · Humanoid | expert behaviors imitated | Helped |
| Add the latent intention cost to the imitation objective (ablation) | Imitation GAN · Reacher, 2 targets | target modes learned | Hurt |
| Add the latent intention cost to the imitation objective (ablation) | Imitation GAN · Walker-2D | expert behaviors imitated | Hurt |
| Add the latent intention cost to the imitation objective (ablation) | Imitation GAN · Humanoid | expert behaviors imitated | Hurt |
| Use continuous rather than categorical intention (ablation) | Multi-modal imitation learning · Reacher, 2 targets | target modes learned | No clear effect |
| Use continuous rather than categorical intention (ablation) | Multi-modal imitation learning · Reacher, 4 targets | target modes learned | Hurt |
| Use multiple intention categories for one demonstrated target behavior (ablation) | Multi-modal imitation learning · Reacher, 1 target | distinct trajectories to target | Helped |
| Use the proposed method on a high-dimensional Humanoid task | Multi-modal imitation learning · Humanoid running backwards | imitation performance | Hurt |
Closest papers by meaning
- InfoGAIL: Interpretable Imitation Learning from Visual Demonstrations (NeurIPS 2017 · similarity 0.89)
- Unlabeled Imperfect Demonstrations in Adversarial Imitation Learning (AAAI 2023 · similarity 0.88)
- Learn what matters: cross-domain imitation learning with task-relevant embeddings (NeurIPS 2022 · similarity 0.88)
- Imitation Learning from Vague Feedback (NeurIPS 2023 · similarity 0.88)
- One-Shot Imitation Learning (NeurIPS 2017 · similarity 0.88)
- Adversarial Imitation Learning with Preferences (ICLR 2023 · 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.