# Learning Disentangled Representations for Time Series

arXiv: https://arxiv.org/abs/2105.08179 · HTML: https://mlsanity.com/paper/2105.08179/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 42 references that are in the map.
- First posted on arXiv 17 May 2021.

## Builds on (most cited first, 27 of 27)

- [Unsupervised Domain Adaptation by Backpropagation](https://mlsanity.com/paper/1409.7495/) (2014)
- [An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling](https://mlsanity.com/paper/1803.01271/) (2018)
- [Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations](https://mlsanity.com/paper/1811.12359/) (2018)
- [Deep Learning and the Information Bottleneck Principle](https://mlsanity.com/paper/1503.02406/) (2015)
- [Deep Domain Confusion: Maximizing for Domain Invariance](https://mlsanity.com/paper/1412.3474/) (2014)
- [InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANs](https://mlsanity.com/paper/2005.09635/) (2020)
- [A Recurrent Latent Variable Model for Sequential Data](https://mlsanity.com/paper/1506.02216/) (2015)
- [Variational Lossy Autoencoder](https://mlsanity.com/paper/1611.02731/) (2016)
- [GP-VAE: Deep Probabilistic Time Series Imputation](https://mlsanity.com/paper/1907.04155/) (2019)
- [Towards Deeper Graph Neural Networks with Differentiable Group Normalization](https://mlsanity.com/paper/2006.06972/) (2020)
- [Multi-Source Deep Domain Adaptation with Weak Supervision for Time-Series Sensor Data](https://mlsanity.com/paper/2005.10996/) (2020)
- [Domain-Adversarial Neural Networks](https://mlsanity.com/paper/1412.4446/) (2014)
- [SOM-VAE: Interpretable Discrete Representation Learning on Time Series](https://mlsanity.com/paper/1806.02199/) (2018)
- [Learning Disentangled Semantic Representation for Domain Adaptation](https://mlsanity.com/paper/2012.11807/) (2020)
- [TODS: An Automated Time Series Outlier Detection System](https://mlsanity.com/paper/2009.09822/) (2020)
- [SpecAE: Spectral AutoEncoder for Anomaly Detection in Attributed Networks](https://mlsanity.com/paper/1908.03849/) (2019)
- [Meta-AAD: Active Anomaly Detection with Deep Reinforcement Learning](https://mlsanity.com/paper/2009.07415/) (2020)
- [A Hybrid Convolutional Variational Autoencoder for Text Generation](https://mlsanity.com/paper/1702.02390/) (2017)
- [Is a Single Vector Enough? Exploring Node Polysemy for Network Embedding](https://mlsanity.com/paper/1905.10668/) (2019)
- [Dual Policy Distillation](https://mlsanity.com/paper/2006.04061/) (2020)
- [Interpretable Deep Graph Generation with Node-Edge Co-Disentanglement](https://mlsanity.com/paper/2006.05385/) (2020)
- [PyODDS: An End-to-end Outlier Detection System with Automated Machine Learning](https://mlsanity.com/paper/2003.05602/) (2020)
- [Towards Generalizable Deepfake Detection with Locality-aware AutoEncoder](https://mlsanity.com/paper/1909.05999/) (2019)
- [Riemannian Normalizing Flow on Variational Wasserstein Autoencoder for Text Modeling](https://mlsanity.com/paper/1904.02399/) (2019)
- [Deep Structured Cross-Modal Anomaly Detection](https://mlsanity.com/paper/1908.03848/) (2019)
- [AutoOD: Automated Outlier Detection via Curiosity-guided Search and Self-imitation Learning](https://mlsanity.com/paper/2006.11321/) (2020)
- [XDeep: An Interpretation Tool for Deep Neural Networks](https://mlsanity.com/paper/1911.01005/) (2019)

## Built on by (most cited first, 8 of 8)

- [Generative Time Series Forecasting with Diffusion, Denoise, and Disentanglement](https://mlsanity.com/paper/2301.03028/) (2023)
- [Diffusion Models for Time Series Applications: A Survey](https://mlsanity.com/paper/2305.00624/) (2023)
- [The Rise of Diffusion Models in Time-Series Forecasting](https://mlsanity.com/paper/2401.03006/) (2024)
- [Learning to Walk with Dual Agents for Knowledge Graph Reasoning](https://mlsanity.com/paper/2112.12876/) (2021)
- [Interpretation of Time-Series Deep Models: A Survey](https://mlsanity.com/paper/2305.14582/) (2023)
- [Towards Similarity-Aware Time-Series Classification](https://mlsanity.com/paper/2201.01413/) (2022)
- [Learning Actionable World Models for Industrial Process Control](https://mlsanity.com/paper/2503.01411/) (2025)
- [Aionoscope: Debugging Latent-State Accessibility in Time-Series Representations](https://mlsanity.com/paper/2607.00956/) (2026)

## Abstract (arXiv)

Time-series representation learning is a fundamental task for time-series analysis. While significant progress has been made to achieve accurate representations for downstream applications, the learned representations often lack interpretability and do not expose semantic meanings. Different from previous efforts on the entangled feature space, we aim to extract the semantic-rich temporal correlations in the latent interpretable factorized representation of the data. Motivated by the success of disentangled representation learning in computer vision, we study the possibility of learning semantic-rich time-series representations, which remains unexplored due to three main challenges: 1) sequential data structure introduces complex temporal correlations and makes the latent representations hard to interpret, 2) sequential models suffer from KL vanishing problem, and 3) interpretable semantic concepts for time-series often rely on multiple factors instead of individuals. To bridge the gap, we propose Disentangle Time Series (DTS), a novel disentanglement enhancement framework for sequential data. Specifically, to generate hierarchical semantic concepts as the interpretable and disentangled representation of time-series, DTS introduces multi-level disentanglement strategies by covering both individual latent factors and group semantic segments. We further theoretically show how to alleviate the KL vanishing problem: DTS introduces a mutual information maximization term, while preserving a heavier penalty on the total correlation and the dimension-wise KL to keep the disentanglement property. Experimental results on various real-world benchmark datasets demonstrate that the representations learned by DTS achieve superior performance in downstream applications, with high interpretability of semantic concepts.