Research map / Time series / Causal and generative time series
Deep Explicit Duration Switching Models for Time Series
Abdul Fatir Ansari, Konstantinos Benidis, Richard Kurle, Ali Caner Turkmen, Harold Soh, Alex Smola, Bernie Wang, Tim Januschowski
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 |
|---|---|---|---|
| Add explicit duration modeling to SNLDS (ablation) | SNLDS · 3 mode system | accuracy: 0.82\textpm0.08 -> 0.97\textpm0.00 | Improved |
| Add explicit duration modeling to SNLDS (ablation) | SNLDS · dancing bees | accuracy: 0.44\textpm0.01 -> 0.56\textpm0.06 | Improved |
| Add state-to-switch recurrence to ED-SDS (ablation) | ED-SDS · bouncing ball | accuracy: 0.95\textpm0.00 -> 0.97\textpm0.00 | Improved |
| Add state-to-switch recurrence to ED-SDS (ablation) | ED-SDS · 3 mode system | accuracy: 0.97\textpm0.00 -> 0.98\textpm0.00 | Improved |
| Add state-to-switch recurrence to ED-SDS (ablation) | ED-SDS · dancing bees | accuracy: 0.56\textpm0.06 -> 0.73\textpm0.10 | Improved |
| Add state-to-switch recurrence to ED-SDS (ablation) | ED-SDS · dancing bees (K=2) | accuracy: 0.79\textpm0.09 -> 0.91\textpm0.04 | Improved |
| Use RED-SDS rather than the strongest reported baseline | time-series forecasting · exchange | CRPS: 0.014\textpm0.001 -> 0.013\textpm0.001 | Improved |
| Use RED-SDS rather than the strongest reported baseline | time-series forecasting · solar | CRPS: 0.358\textpm0.001 -> 0.419\textpm0.010 | Hurt |
| Use RED-SDS rather than the strongest reported baseline | time-series forecasting · electricity | CRPS: 0.062\textpm0.004 -> 0.066\textpm0.002 | Hurt |
| Use RED-SDS rather than the strongest reported baseline | time-series forecasting · traffic | CRPS: 0.131\textpm0.005 -> 0.129\textpm0.002 | Improved |
| Use RED-SDS rather than the strongest reported baseline | time-series forecasting · wiki | CRPS: 0.283\textpm 0.006 -> 0.318\textpm0.006 | Hurt |
Closest papers by meaning
- Time series analysis with Gumbel dynamics (NeurIPS 2026 · similarity 0.88)
- End-to-End Identifiable and Consistent Recurrent Switching Dynamical Systems (NeurIPS 2026 · similarity 0.87)
- On the Identifiability of Switching Dynamical Systems (ICML 2024 · similarity 0.86)
- Inference of Neural Dynamics Using Switching Recurrent Neural Networks (NeurIPS 2024 · similarity 0.86)
- Modeling Latent Neural Dynamics with Gaussian Process Switching Linear Dynamical Systems (NeurIPS 2024 · similarity 0.84)
- Amortized Control of Continuous State Space Feynman-Kac Model for Irregular Time Series (ICLR 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.
