Research map / Reinforcement learning / Imitation learning
Imitation Learning via Kernel Mean Embedding
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 |
|---|---|---|---|
| replace GAIL discriminator (JS divergence) with MMD witness-function cost (GMMIL) | imitation learning with TRPO · MuJoCo tasks with scarce expert trajectories (Reacher, Walker, Ant) | scaled return: up to 41% gain | Helped |
| replace GAIL with GMMIL when expert data is plentiful | imitation learning with TRPO · OpenAI Gym control tasks (largest expert set) | scaled return | No clear effect |
| replace GAIL with GMMIL | imitation learning with TRPO · MuJoCo tasks, iterations to converge | sample efficiency (iterations) | Helped |
| perform 5 TRPO updates per iteration with importance sampling (ablation) | GMMIL · high-dimensional MuJoCo tasks | convergence rate | Mixed |
| perform 5 TRPO updates per iteration with importance sampling | GAIL · high-dimensional MuJoCo tasks | training progress | Hurt |
| vary the discriminator learning rate away from the best value | GAIL · Reacher, HalfCheetah, Hopper, Walker | scaled return | Hurt |
| use two Gaussian kernels with median-heuristic bandwidths | GMMIL · all tasks | imitation success | Helped |
| use time-discounted sum of per-step distributions instead of stationary unnormalized distribution (ablation) | GMMIL · control tasks | performance | No clear effect |
Closest papers by meaning
- Variational Adversarial Kernel Learned Imitation Learning (AAAI 2020 · similarity 0.93)
- Generative Adversarial Imitation Learning (NeurIPS 2016 · similarity 0.92)
- A Bayesian Approach to Generative Adversarial Imitation Learning (NeurIPS 2018 · similarity 0.92)
- Exploring Gradient Explosion in Generative Adversarial Imitation Learning: A Probabilistic Perspective (AAAI 2024 · similarity 0.90)
- Diffusing States and Matching Scores: A New Framework for Imitation Learning (ICLR 2025 · similarity 0.90)
- Diffusion Imitation from Observation (NeurIPS 2024 · similarity 0.90)
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.