Research map / Probabilistic and scientific modeling / Causal discovery
Maximum Causal Tsallis Entropy Imitation Learning
Kyungjae Lee, Sungjoon Choi, Songhwai Oh
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 MCTEIL with a sparse MDN instead of soft GAIL with a soft MDN | Multi-goal environment · Multi-goal environment | reachability | Improved |
| Use MCTEIL with a sparse MDN instead of soft GAIL with a soft MDN | Multi-goal environment · Multi-goal environment | average return | No clear effect |
| Use MCTEIL instead of BC, GAIL, or soft GAIL | Imitation learning methods · MuJoCo continuous control | average return | Improved |
| Use MCTEIL instead of soft GAIL | Imitation learning methods · Walker2d | average return | No clear effect |
| Use MCTEIL instead of BC or GAIL | Imitation learning methods · Reacher | average return | Improved |
| Replace a single Gaussian policy with an MDN policy | Imitation learning policy · MuJoCo continuous control | average return | Improved |
| Replace sparsemax mixture weights with softmax mixture weights (ablation) | MCTEIL sparse MDN · Multi-goal environment | reachability | Hurt |
| Replace sparsemax mixture weights with softmax mixture weights (ablation) | MCTEIL sparse MDN · MuJoCo continuous control | average return | Hurt |
| Replace the multi-modal MDN with a single Gaussian (ablation) | Imitation learning policy · Multi-goal environment | performance | Hurt |
| Estimate Tsallis entropy using sampled action likelihoods instead of analytically integrating the mixture density (ablation) | MCTEIL entropy estimate · Imitation learning | performance | Hurt |
Closest papers by meaning
- Causal Imitation for Markov Decision Processes: a Partial Identification Approach (NeurIPS 2024 · similarity 0.85)
- DIME: Diffusion-Based Maximum Entropy Reinforcement Learning (ICML 2025 · similarity 0.85)
- Imitation from Observations with Trajectory-Level Generative Embeddings (NeurIPS 2026 · similarity 0.84)
- Imitation Learning from Purified Demonstrations (ICML 2024 · similarity 0.83)
- Beyond-Expert Performance with Limited Demonstrations: Efficient Imitation Learning with Double Exploration (ICLR 2025 · similarity 0.83)
- Maximum Entropy Reinforcement Learning with Diffusion Policy (ICML 2025 · similarity 0.83)
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.
