Causal Discovery from Conditionally Stationary Time Series

Carles Balsells-Rodas, Xavier Sumba, Tanmayee Narendra, Ruibo Tu, Gabriele Schweikert, Hedvig Kjellstrom, Yingzhen Li
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:2715-2741, 2025.

Abstract

Causal discovery, i.e., inferring underlying causal relationships from observational data, is highly challenging for AI systems. In a time series modeling context, traditional causal discovery methods mainly consider constrained scenarios with fully observed variables and/or data from stationary time-series. We develop a causal discovery approach to handle a wide class of nonstationary time series that are conditionally stationary, where the nonstationary behaviour is modeled as stationarity conditioned on a set of latent state variables. Named State-Dependent Causal Inference (SDCI), our approach is able to recover the underlying causal dependencies, with provable identifiablity for the state-dependent causal structures. Empirical experiments on nonlinear particle interaction data and gene regulatory networks demonstrate SDCI’s superior performance over baseline causal discovery methods. Improved results over non-causal RNNs on modeling NBA player movements demonstrate the potential of our method and motivate the use of causality-driven methods for forecasting.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-balsells-rodas25a, title = {Causal Discovery from Conditionally Stationary Time Series}, author = {Balsells-Rodas, Carles and Sumba, Xavier and Narendra, Tanmayee and Tu, Ruibo and Schweikert, Gabriele and Kjellstrom, Hedvig and Li, Yingzhen}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {2715--2741}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/balsells-rodas25a/balsells-rodas25a.pdf}, url = {https://proceedings.mlr.press/v267/balsells-rodas25a.html}, abstract = {Causal discovery, i.e., inferring underlying causal relationships from observational data, is highly challenging for AI systems. In a time series modeling context, traditional causal discovery methods mainly consider constrained scenarios with fully observed variables and/or data from stationary time-series. We develop a causal discovery approach to handle a wide class of nonstationary time series that are conditionally stationary, where the nonstationary behaviour is modeled as stationarity conditioned on a set of latent state variables. Named State-Dependent Causal Inference (SDCI), our approach is able to recover the underlying causal dependencies, with provable identifiablity for the state-dependent causal structures. Empirical experiments on nonlinear particle interaction data and gene regulatory networks demonstrate SDCI’s superior performance over baseline causal discovery methods. Improved results over non-causal RNNs on modeling NBA player movements demonstrate the potential of our method and motivate the use of causality-driven methods for forecasting.} }
Endnote
%0 Conference Paper %T Causal Discovery from Conditionally Stationary Time Series %A Carles Balsells-Rodas %A Xavier Sumba %A Tanmayee Narendra %A Ruibo Tu %A Gabriele Schweikert %A Hedvig Kjellstrom %A Yingzhen Li %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-balsells-rodas25a %I PMLR %P 2715--2741 %U https://proceedings.mlr.press/v267/balsells-rodas25a.html %V 267 %X Causal discovery, i.e., inferring underlying causal relationships from observational data, is highly challenging for AI systems. In a time series modeling context, traditional causal discovery methods mainly consider constrained scenarios with fully observed variables and/or data from stationary time-series. We develop a causal discovery approach to handle a wide class of nonstationary time series that are conditionally stationary, where the nonstationary behaviour is modeled as stationarity conditioned on a set of latent state variables. Named State-Dependent Causal Inference (SDCI), our approach is able to recover the underlying causal dependencies, with provable identifiablity for the state-dependent causal structures. Empirical experiments on nonlinear particle interaction data and gene regulatory networks demonstrate SDCI’s superior performance over baseline causal discovery methods. Improved results over non-causal RNNs on modeling NBA player movements demonstrate the potential of our method and motivate the use of causality-driven methods for forecasting.
APA
Balsells-Rodas, C., Sumba, X., Narendra, T., Tu, R., Schweikert, G., Kjellstrom, H. & Li, Y.. (2025). Causal Discovery from Conditionally Stationary Time Series. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:2715-2741 Available from https://proceedings.mlr.press/v267/balsells-rodas25a.html.

Related Material