Research map / Probabilistic and scientific modeling / Dynamical systems
Switching Autoregressive Low-rank Tensor Models
Hyun Dong Lee, Andrew Warrington, Joshua I Glaser, Scott Linderman
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 low-rank tensor factors for the ARHMM dynamics | CP-SALT and Tucker-SALT · Mouse behavior | held-out test log-likelihood | Improved |
| Increase the number of lags in the autoregressive model | ARHMM · Mouse behavior | held-out validation log-likelihood | Hurt |
| Use low-rank tensor factorization instead of unrestricted autoregressive tensors | CP-SALT · Mouse behavior | overfitting as lags increase | Improved |
| Use CP-SALT instead of SLDS or ARHMM | C. elegans neural data · C. elegans neural data | held-out explained variance ratio | No clear effect |
| Use CP-SALT instead of ARHMM | C. elegans neural data · C. elegans neural data | parameter count | Improved |
| Use rank 10 for CP-SALT or rank 7 for Tucker-SALT (ablation) | SALT · Simulated LDS | autoregressive tensor reconstruction mean squared error | Improved |
| Use rank 7 rather than a larger rank (ablation) | CP-SALT and Tucker-SALT · Simulated LDS | held-out test log-likelihood | Improved |
| Use more autoregressive lags (ablation) | ARHMM · Simulated LDS | data required to fit | Hurt |
| Use SALT to approximate a recurrent switching LDS | SALT · Synthetic recurrent SLDS | latent trajectory and state segmentation recovery | Improved |
| Use SALT to model nonlinear Lorenz dynamics | SALT · Synthetic Lorenz attractor | latent and observation reconstruction | Improved |
| Use SALT rather than TVART | Synthetic switching LDS data · Synthetic switching LDS | autoregressive dynamics tensor reconstruction | Improved |
| Use fewer discrete states in the learned segmentation | CP-SALT · Mouse behavior | segmentation parsimony | Improved |
Closest papers by meaning
- Modeling Latent Neural Dynamics with Gaussian Process Switching Linear Dynamical Systems (NeurIPS 2024 · similarity 0.87)
- Probabilistic Decomposed Linear Dynamical Systems for Robust Discovery of Latent Neural Dynamics (NeurIPS 2024 · similarity 0.86)
- On the Identifiability of Switching Dynamical Systems (ICML 2024 · similarity 0.86)
- Time series analysis with Gumbel dynamics (NeurIPS 2026 · similarity 0.85)
- Modeling Latent Non-Linear Dynamical System over Time Series (AAAI 2025 · similarity 0.85)
- Inferring stochastic low-rank recurrent neural networks from neural data (NeurIPS 2024 · similarity 0.85)
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.
