Learning domain-specific causal discovery from time series
This work addresses the challenge of improving causal discovery accuracy in domains like neuroscience and medicine, offering a new supervised approach that could benefit the broader machine learning community, though it appears incremental as it builds on existing data-driven paradigms.
The paper tackled the problem of causal discovery from time series data by proposing a data-driven approach that learns domain-specific procedures from extensive datasets with known causal relationships, which significantly outperformed human-designed methods like Mutual Information, VAR-LiNGAM, and Granger Causality on datasets including the MOS 6502 microprocessor, NetSim fMRI, and Dream3 gene datasets.
Causal discovery (CD) from time-varying data is important in neuroscience, medicine, and machine learning. Techniques for CD encompass randomized experiments, which are generally unbiased but expensive, and algorithms such as Granger causality, conditional-independence-based, structural-equation-based, and score-based methods that are only accurate under strong assumptions made by human designers. However, as demonstrated in other areas of machine learning, human expertise is often not entirely accurate and tends to be outperformed in domains with abundant data. In this study, we examine whether we can enhance domain-specific causal discovery for time series using a data-driven approach. Our findings indicate that this procedure significantly outperforms human-designed, domain-agnostic causal discovery methods, such as Mutual Information, VAR-LiNGAM, and Granger Causality on the MOS 6502 microprocessor, the NetSim fMRI dataset, and the Dream3 gene dataset. We argue that, when feasible, the causality field should consider a supervised approach in which domain-specific CD procedures are learned from extensive datasets with known causal relationships, rather than being designed by human specialists. Our findings promise a new approach toward improving CD in neural and medical data and for the broader machine learning community.