Research map / Probabilistic and scientific modeling / Variational and generative sampling
Learning Predictive State Representations From Non-Uniform Sampling
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 |
|---|---|---|---|
| Add denoising (weighted Frobenius norm with sample-count weights) to spectral PSR learning | Spectral PSR learning · One-state environment, k=100 | SDM matrix error | Helped |
| Add denoising to spectral PSR learning | Spectral PSR learning · One-state environment, k=100 | one-step prediction error | Helped |
| Add denoising to spectral PSR learning | Spectral PSR learning · Torus environment (4 and 20 observations) | prediction error and SDM matrix error | Helped |
| Use denoised PSR model for state representation in planning | Fitted-Q iteration on learned PSR · Torus planning | average total reward | Helped |
| Give standard spectral learning more training data in the policy-constrained torus | Standard spectral PSR learning · Torus planning | planning reward | Hurt |
| Add denoising to spectral PSR learning | Spectral PSR learning · Robot sensing (24 sensors, rank-40 SDM) | test log-likelihood | Helped |
| Increase SDM rank beyond 40 | Both PSR methods · Robot sensing | log-likelihood | No clear effect |
| Vary first penalty weight lambda1 (10 vs 0.005) with lambda2 = 0 (ablation) | Denoising spectral learning · Torus, 4 observations | prediction error | Mixed |
Closest papers by meaning
- Learning invariant representations of time-homogeneous stochastic dynamical systems (ICLR 2024 · similarity 0.86)
- Identifying latent state transitions in non-linear dynamical systems (ICLR 2025 · similarity 0.85)
- Predicting partially observable dynamical systems via diffusion models with a multiscale inference scheme (NeurIPS 2025 · similarity 0.85)
- On conditional diffusion models for PDE simulations (NeurIPS 2024 · similarity 0.84)
- Learning Hybrid Dynamics Models with Simulator-Informed Latent States (AAAI 2024 · similarity 0.84)
- Spectral Convolutional Conditional Neural Process (NeurIPS 2025 · similarity 0.84)
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.