Research map / Reinforcement learning / MDPs and planning
Wasserstein Auto-encoded MDPs: Formal Verification of Efficiently Distilled RL Policies with Many-sided Guarantees
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 W²AE-MDP instead of VAE-MDP for policy distillation | RL policy distillation · OpenAI benchmark environments | distillation training speed: up to 10 times faster | Helped |
| Use W²AE-MDP instead of VAE-MDP | latent model learning · OpenAI benchmark environments | abstraction quality | Helped |
| Use W²AE-MDP instead of VAE-MDP | policy distillation · OpenAI benchmark environments | episode return convergence | Helped |
| Use the learned W²AE-MDP transition model without a posterior reconstruction step | W²AE-MDP · OpenAI benchmark environments | PAC local-loss bounds | Helped |
| Distill the original policy into a discrete latent policy | W²AE-MDP · OpenAI benchmark environments | episode return | Helped |
| Use W²AE-MDP instead of VAE-MDP when distilled policies have comparable performance | latent model verification · OpenAI benchmark environments | PAC local-loss bounds | Mixed |
| Use W²AE-MDP rather than a VAE-based latent model | latent representation learning · OpenAI benchmark environments | posterior collapse | Helped |
Closest papers by meaning
- Distillation of RL Policies with Formal Guarantees via Variational Abstraction of Markov Decision Processes (AAAI 2022 · similarity 0.93)
- Transferable Reinforcement Learning via Probabilistic Latent Embeddings and Dynamic Policy Adaptation for Sim-to-Real Deployment (ICML 2026 · similarity 0.88)
- Near-Optimal Sample Complexity for Online Constrained MDPs (NeurIPS 2025 · similarity 0.87)
- Variational Delayed Policy Optimization (NeurIPS 2024 · similarity 0.87)
- Oracle-Efficient Reinforcement Learning for Max Value Ensembles (NeurIPS 2024 · similarity 0.86)
- Provably Efficient CVaR RL in Low-rank MDPs (ICLR 2024 · similarity 0.86)
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.