Research map / Reinforcement learning / Imitation learning
A Bayesian Approach to Generative Adversarial Imitation Learning
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 |
|---|---|---|---|
| use Bayesian predictive cost (SVGD, 5 discriminator particles) instead of point-estimate discriminator | tuned GAIL · Ant-v1 | average return: 4759.12 -> 4808.90 | Helped |
| use Bayesian predictive cost instead of point-estimate discriminator | tuned GAIL · Hopper-v1 | average return: 3595.30 -> 3613.94 | Helped |
| use Bayesian predictive cost instead of point-estimate discriminator | tuned GAIL · HalfCheetah-v1 | average return: 5022.93 -> 4970.77 | No clear effect |
| use Bayesian predictive cost instead of point-estimate discriminator | tuned GAIL · Humanoid-v1 | average return: 10329.66 -> 10388.34 | No clear effect |
| use Bayesian predictive cost with few samples per iteration (1000 or 5000 pairs) | tuned GAIL · high-dimensional MuJoCo tasks sample efficiency | convergence speed | Helped |
| use Bayesian predictive cost with few samples per iteration | tuned GAIL · Hopper-v1 sample efficiency | convergence speed | No clear effect |
| tune GAIL (state-dependent variance, 5 discriminator updates, full trajectories) | GAIL · Ant-v1 | average return: 4132.90 -> 4759.12 | Helped |
Closest papers by meaning
- Exploring Gradient Explosion in Generative Adversarial Imitation Learning: A Probabilistic Perspective (AAAI 2024 · similarity 0.93)
- On Computation and Generalization of Generative Adversarial Imitation Learning (ICLR 2020 · similarity 0.92)
- Imitation Learning via Kernel Mean Embedding (AAAI 2018 · similarity 0.92)
- C-GAIL: Stabilizing Generative Adversarial Imitation Learning with Control Theory (NeurIPS 2024 · similarity 0.90)
- Diffusing States and Matching Scores: A New Framework for Imitation Learning (ICLR 2025 · similarity 0.89)
- Diffusion-Reward Adversarial Imitation Learning (NeurIPS 2024 · 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.