Cause-effect inference through spectral independence in linear dynamical systems: theoretical foundations

Michel Besserve, Naji Shajarisales, Dominik Janzing, Bernhard Schölkopf
Proceedings of the First Conference on Causal Learning and Reasoning, PMLR 177:110-143, 2022.

Abstract

Distinguishing between cause and effect using time series observational data is a major challenge in many scientific fields. A new perspective has been provided based on the principle of Independence of Causal Mechanisms (ICM), leading to the Spectral Independence Criterion (SIC) for time series causally unidirectionally linked by a linear time-invariant relation. SIC postulates that the power spectral density (PSD) of the cause time series is {\it uncorrelated} with the squared modulus of the frequency response of the filter generating the effect. Since SIC rests on methods and assumptions in stark contrast with most causal discovery methods for time series, it raises questions regarding what theoretical grounds justify its use. In this paper, we provide answers covering several key aspects. After providing an information theoretic interpretation of SIC, we present an identifiability result that sheds light on the context for which this approach is expected to perform well. We further demonstrate the robustness of SIC to downsampling – an obstacle that can spoil Granger-based inference. Finally, an invariance perspective allows to explore the limitations of the spectral independence assumption and how to generalize it. Overall, these results provide insights on how the ICM principle can be assessed mathematically to infer direction of causation in empirical time series.

Cite this Paper

BibTeX
@InProceedings{pmlr-v177-besserve22a, title = {Cause-effect inference through spectral independence in linear dynamical systems: theoretical foundations}, author = {Besserve, Michel and Shajarisales, Naji and Janzing, Dominik and Sch{\"o}lkopf, Bernhard}, booktitle = {Proceedings of the First Conference on Causal Learning and Reasoning}, pages = {110--143}, year = {2022}, editor = {Schölkopf, Bernhard and Uhler, Caroline and Zhang, Kun}, volume = {177}, series = {Proceedings of Machine Learning Research}, month = {11--13 Apr}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v177/besserve22a/besserve22a.pdf}, url = {https://proceedings.mlr.press/v177/besserve22a.html}, abstract = {Distinguishing between cause and effect using time series observational data is a major challenge in many scientific fields. A new perspective has been provided based on the principle of Independence of Causal Mechanisms (ICM), leading to the Spectral Independence Criterion (SIC) for time series causally unidirectionally linked by a linear time-invariant relation. SIC postulates that the power spectral density (PSD) of the cause time series is {\it uncorrelated} with the squared modulus of the frequency response of the filter generating the effect. Since SIC rests on methods and assumptions in stark contrast with most causal discovery methods for time series, it raises questions regarding what theoretical grounds justify its use. In this paper, we provide answers covering several key aspects. After providing an information theoretic interpretation of SIC, we present an identifiability result that sheds light on the context for which this approach is expected to perform well. We further demonstrate the robustness of SIC to downsampling – an obstacle that can spoil Granger-based inference. Finally, an invariance perspective allows to explore the limitations of the spectral independence assumption and how to generalize it. Overall, these results provide insights on how the ICM principle can be assessed mathematically to infer direction of causation in empirical time series.} }
Endnote
%0 Conference Paper %T Cause-effect inference through spectral independence in linear dynamical systems: theoretical foundations %A Michel Besserve %A Naji Shajarisales %A Dominik Janzing %A Bernhard Schölkopf %B Proceedings of the First Conference on Causal Learning and Reasoning %C Proceedings of Machine Learning Research %D 2022 %E Bernhard Schölkopf %E Caroline Uhler %E Kun Zhang %F pmlr-v177-besserve22a %I PMLR %P 110--143 %U https://proceedings.mlr.press/v177/besserve22a.html %V 177 %X Distinguishing between cause and effect using time series observational data is a major challenge in many scientific fields. A new perspective has been provided based on the principle of Independence of Causal Mechanisms (ICM), leading to the Spectral Independence Criterion (SIC) for time series causally unidirectionally linked by a linear time-invariant relation. SIC postulates that the power spectral density (PSD) of the cause time series is {\it uncorrelated} with the squared modulus of the frequency response of the filter generating the effect. Since SIC rests on methods and assumptions in stark contrast with most causal discovery methods for time series, it raises questions regarding what theoretical grounds justify its use. In this paper, we provide answers covering several key aspects. After providing an information theoretic interpretation of SIC, we present an identifiability result that sheds light on the context for which this approach is expected to perform well. We further demonstrate the robustness of SIC to downsampling – an obstacle that can spoil Granger-based inference. Finally, an invariance perspective allows to explore the limitations of the spectral independence assumption and how to generalize it. Overall, these results provide insights on how the ICM principle can be assessed mathematically to infer direction of causation in empirical time series.
APA
Besserve, M., Shajarisales, N., Janzing, D. & Schölkopf, B.. (2022). Cause-effect inference through spectral independence in linear dynamical systems: theoretical foundations. Proceedings of the First Conference on Causal Learning and Reasoning, in Proceedings of Machine Learning Research 177:110-143 Available from https://proceedings.mlr.press/v177/besserve22a.html.

Related Material