# Deep Explicit Duration Switching Models for Time Series

arXiv: https://arxiv.org/abs/2110.13878 · HTML: https://mlsanity.com/paper/2110.13878/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 50 references that are in the map.
- First posted on arXiv 26 Oct 2021.

## Builds on (most cited first, 12 of 12)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networks](https://mlsanity.com/paper/1704.04110/) (2017)
- [Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks](https://mlsanity.com/paper/1703.07015/) (2017)
- [Composing graphical models with neural networks for structured representations and fast inference](https://mlsanity.com/paper/1603.06277/) (2016)
- [A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learning](https://mlsanity.com/paper/1710.05741/) (2017)
- [Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks](https://mlsanity.com/paper/1301.3853/) (2013)
- [Variational Temporal Abstraction](https://mlsanity.com/paper/1910.00775/) (2019)
- [Recurrent switching linear dynamical systems](https://mlsanity.com/paper/1610.08466/) (2016)
- [Collapsed Amortized Variational Inference for Switching Nonlinear Dynamical Systems](https://mlsanity.com/paper/1910.09588/) (2019)
- [Switching Linear Dynamics for Variational Bayes Filtering](https://mlsanity.com/paper/1905.12434/) (2019)
- [Explicit-Duration Markov Switching Models](https://mlsanity.com/paper/1909.05800/) (2019)
- [Bayesian Online Prediction of Change Points](https://mlsanity.com/paper/1902.04524/) (2019)

## Built on by (most cited first, 17 of 17)

- [Chronos: Learning the Language of Time Series](https://mlsanity.com/paper/2403.07815/) (2024)
- [Neural Continuous-Discrete State Space Models for Irregularly-Sampled Time Series](https://mlsanity.com/paper/2301.11308/) (2023)
- [Learning Mixtures of Linear Dynamical Systems](https://mlsanity.com/paper/2201.11211/) (2022)
- [Graph Switching Dynamical Systems](https://mlsanity.com/paper/2306.00370/) (2023)
- [On the Identifiability of Switching Dynamical Systems](https://mlsanity.com/paper/2305.15925/) (2023)
- [Continuous Locomotive Crowd Behavior Generation](https://mlsanity.com/paper/2504.04756/) (2025)
- [Conformal Prediction for Time-series Forecasting with Change Points](https://mlsanity.com/paper/2509.02844/) (2025)
- [T-SaS: Toward Shift-aware Dynamic Adaptation for Streaming Data](https://mlsanity.com/paper/2309.02610/) (2023)
- [Amortized Equation Discovery in Hybrid Dynamical Systems](https://mlsanity.com/paper/2406.03818/) (2024)
- [DynaConF: Dynamic Forecasting of Non-Stationary Time Series](https://mlsanity.com/paper/2209.08411/) (2022)
- [Discovering group dynamics in coordinated time series via hierarchical recurrent switching-state models](https://mlsanity.com/paper/2401.14973/) (2024)
- [Towards Dynamic Trend Filtering through Trend Point Detection with Reinforcement Learning](https://mlsanity.com/paper/2406.03665/) (2024)
- [MBDS: A Multi-Body Dynamics Simulation Dataset for Graph Networks Simulators](https://mlsanity.com/paper/2410.03107/) (2024)
- [End-to-End Identifiable and Consistent Recurrent Switching Dynamical Systems](https://mlsanity.com/paper/2605.06315/) (2026)
- [WaveBound: Dynamic Error Bounds for Stable Time Series Forecasting](https://mlsanity.com/paper/2210.14303/) (2022)
- [Bayesian Inference in Recurrent Explicit Duration Switching Linear Dynamical Systems](https://mlsanity.com/paper/2411.04280/) (2024)
- [Diffusion-Driven State Space Models](https://mlsanity.com/paper/2606.21036/) (2026)

## Abstract (arXiv)

Many complex time series can be effectively subdivided into distinct regimes that exhibit persistent dynamics. Discovering the switching behavior and the statistical patterns in these regimes is important for understanding the underlying dynamical system. We propose the Recurrent Explicit Duration Switching Dynamical System (RED-SDS), a flexible model that is capable of identifying both state- and time-dependent switching dynamics. State-dependent switching is enabled by a recurrent state-to-switch connection and an explicit duration count variable is used to improve the time-dependent switching behavior. We demonstrate how to perform efficient inference using a hybrid algorithm that approximates the posterior of the continuous states via an inference network and performs exact inference for the discrete switches and counts. The model is trained by maximizing a Monte Carlo lower bound of the marginal log-likelihood that can be computed efficiently as a byproduct of the inference routine. Empirical results on multiple datasets demonstrate that RED-SDS achieves considerable improvement in time series segmentation and competitive forecasting performance against the state of the art.