Research map / Probabilistic and scientific modeling / Dynamical systems
Extracting Latent State Representations with Linear Dynamics from Rich Observations
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 the learned representation with TRPO to train a linear policy after more than 100 thousand representation-learning samples | TRPO with learned representation · Pendulum-v0 | policy performance | Helped |
| Use the learned representation with TRPO after training the representation on 35,000 samples | TRPO with learned representation · Pendulum-v0 | policy performance | Helped |
| Use learned representations and TRPO instead of baseline 2 or Dreamer | Pendulum-v0 · Pendulum-v0 | samples to reach similar performance: around 3 times | Hurt |
| Use the learned representation with TRPO after training the representation on 25,000 samples | TRPO with learned representation · MountainCarContinuous-v0 | reward | Helped |
| Use the learned representation with TRPO after training the representation on 25,000 samples | TRPO with learned representation · MountainCarContinuous-v0 | sample efficiency | No clear effect |
| Use Dreamer | MountainCarContinuous-v0 · MountainCarContinuous-v0 | policy performance | Hurt |
| Use learned high-dimensional representations instead of the original low-dimensional state variables | TRPO with learned representation · MountainCarContinuous-v0 | total reward | Hurt |
Closest papers by meaning
- Rich-Observation Reinforcement Learning with Continuous Latent Dynamics (ICML 2024 · similarity 0.89)
- Identifying latent state transitions in non-linear dynamical systems (ICLR 2025 · similarity 0.87)
- Model-free Low-Rank Reinforcement Learning via Leveraged Entry-wise Matrix Estimation (NeurIPS 2024 · similarity 0.87)
- Reinforcement Learning Under Latent Dynamics: Toward Statistical and Algorithmic Modularity (NeurIPS 2024 · similarity 0.87)
- Identifying Policy Gradient Subspaces (ICLR 2024 · similarity 0.87)
- Learning Low-dimensional Latent Dynamics from High-dimensional Observations: Non-asymptotics and Lower Bounds (ICML 2024 · similarity 0.87)
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.