Causal structures learned from incomplete noisy time series data

Causal Heteroscedastic Structure Learning From Incomplete Temporal data

Machine LearningArtificial Intelligence

Summary

Finding cause-and-effect relationships in data that changes over time is very important but hard, especially when some data is missing or noisy. The authors show that trying to fix missing data and find causes separately does not work well. They introduce a new method called Cheesefill that fixes missing data while learning the causes at the same time, even when the noise changes in different ways. Their method uses advanced math and neural networks to do this optimally. Tests show Cheesefill works better than previous methods on both fake and real data.

What this means in practice

  • For financial analysts: Improve interpretation of incomplete and noisy financial time series to better understand causal risks and opportunities.
  • For neuroscience data teams: Enable more accurate recovery of brain activity causal patterns from partially missing neural recordings with fluctuating noise levels.

Authors

Abdellah Rahmani, Pascal Frossard

Abstract

Discovering causal structures in multivariate time series (MTS) is critical in domains such as finance and neuroscience, yet remains challenging in practice due to three distinct obstacles. First, MTS are routinely corrupted by missing values, where naive impute-then-discover pipelines distort the underlying data distribution and yield spurious causal graphs. Second, identifying instantaneous effects is difficult even on complete data, and missingness further entangles imputation with graph recovery. Third, real-world data frequently exhibits heteroscedastic noise with variance driven by both instantaneous and lagged causes, violating the core assumptions of standard discovery methods. Existing approaches fail to address these challenges concurrently, typically presuming complete data, homoscedastic noise, or strictly time-lagged relationships. Crucially, we show through ablation studies that resolving these issues sequentially, rather than jointly, leads to a substantial degradation in performance. To overcome this, we introduce Cheesefill, an optimal transport framework for causal structure learning under missingness and heteroscedastic noise. Cheesefill parameterizes causal mechanisms using conditional normalizing flows and jointly learns a stochastic correction map that refines naive imputations into trajectories consistent with the inferred causal mechanisms. Theoretically, we prove that optimizing over stochastic correction maps is equivalent to a conditional kernel reformulation of the Kantorovich problem, from which our neural parameterization yields a tractable objective. Extensive experiments on synthetic and real-world benchmarks demonstrate that Cheesefill consistently outperforms existing baselines across diverse missingness regimes.