Machine learning & Causal discovery

A standard maxim is that “correlation is not causation”: we ought not infer a causal relationship between variables simply because they are associated. However, this maxim obscures the ways that patterns of correlation can imply patterns of causation. My research has focused on developing, testing, and validating (when possible) novel causal discovery algorithms for “non-standard” observational and experimental data. A significant part of that work has focused on time series data as part of a long collaboration with Sergey Plis. I have also engaged in multiple collaborations to apply these and other machine learning algorithms to real-world problems across a wide range of domains.

Selected publications:

• Abavisani, M., Danks, D., & Plis, S. (2023). GRACE-C: Generalized rate agnostic causal estimation via constraints. Proceedings of ICLR.

• Schnetz, M., Danks, D., & Mahajan, A. (2023). Preoperative identification of patient-dependent blood pressure targets associated with low risk of intraoperative hypotension during noncardiac surgery. Anesthesia & Analgesia, 136(2): 194-203.

• Solovyeva, K., Danks, D., Abavisani, M., & Plis, S. (2023). Causal learning through deliberate undersampling. Proceedings of CLeaR conference.

• Schnetz, M. P., Hochheiser, H. S., Danks, D. J., Landsittel, D. P., Vogt, K. M., Ibinson, J. W., Whitehurst, S. L., McDermott, S. P., Duque, M. G., & Kaynar, A. M. (2019). The Triple Variable Index combines information generated over time from common monitoring variables to identify patients expressing distinct patterns of intraoperative physiology. BMC Medical Research Methodology, 19(1): 17 pages.

• Malinsky, D., & Danks, D. (2018). Causal discovery algorithms: A practical guide. Philosophy Compass, 13, e12470. doi:10.1111/phc3.12470

• Broger, T., Roy, R. B., Filomena, A., Greef, C. H., Rimmele, S., Havumaki, J., Danks, D., & 11 additional authors. (2017). Diagnostic performance of tuberculosis-specific IgG antibody profiles in patients with presumptive TB from two continents. Serology.

• Hyttinen, A., Plis, S., Jarvisalo, M., Eberhardt, F., & Danks, D. (2017). A constraint optimization approach to causal discovery from subsampled time series data. International Journal of Approximate Reasoning, 90, 208-225.

• Kazman, R., Stoddard, R., Danks, D., & Cai, Y. (2017). Causal modeling, discovery, & inference for software engineering. In Proceedings of 39th International Conference on Software Engineering (ICSE 2017) (pp. 172-174). Piscataway, NJ: IEEE Press.

• Plis, S., Danks, D., Freeman, C., & Calhoun, V. (2015). Rate-agnostic (causal) structure learning. In Advances in neural information processing systems 28 (pp. 3303-3311). La Jolla, CA: The NIPS Foundation. (First two authors contributed equally.)

• Plis, S., Danks, D., & Yang, J. (2015). Mesochronal structure learning. In M. Meila & T. Heskes (Eds.), Uncertainty in artificial intelligence (UAI) 31 (pp. 702-711). Corvallis, OR: AUAI Press. (First two authors contributed equally.)

• Danks, D., & Plis, S. (2014). Learning causal structure from undersampled time series. In JMLR: Workshop and Conference Proceedings. (First two authors contributed equally.)

• Danks, D. (2014). Learning. In K. Frankish & W. Ramsey (Eds.), Cambridge handbook to artificial intelligence (pp. 151-167). Cambridge: Cambridge University Press.

• Kummerfeld, E., & Danks, D. (2013). Tracking time-varying graphical structure. In C.J.C. Burges, L. Bottou, M. Welling, Z. Ghahramani, & K.Q. Weinberger (Eds.), Advances in neural information processing systems 26. La Jolla, CA: The NIPS Foundation.

• Ramapriyan, H., Isaac, D., Yang, W., Bonnlander, B., & Danks, D. (2010). An intelligent archive testbed incorporating data mining. In L. Di & H. K. Ramapriyan (Eds.), Standard-based data and information systems for earth observations (pp. 165-188). Berlin: Springer-Verlag.

• Wimberly, F., Danks, D., Glymour, C., & Chu, T. (2009). Problems for structure learning: Aggregation and computational complexity. In S. Das, D. Caragea, W. H. Hsu, & S. M. Welch (Eds.), Computational methodologies in gene regulatory networks (pp. 310-332). Hershey, PA: IGI Global Publishing.

• Tillman, R. E., Danks, D., & Glymour, C. (2008). Integrating locally learned causal structures with overlapping variables. in D. Koller, D. Schuurmans, Y. Bengio, & L. Bottou (Eds.), Advances in neural information processing systems 21 (pp. 1665-1672). La Jolla, CA: The NIPS Foundation.

• Townsend, K. A., Wollstein, G., Danks, D., Sung, K. R., Ishikawa, H., Kagemann, L., Gabriele, M. L., & Schuman, J. S. (2008). Heidelberg Retina Tomography III machine learning classifiers for glaucoma detection. British Journal of Ophthalmology, 92, 814-818.

• Danks, D. (2002). Learning the causal structure of overlapping variable sets. In S. Lange, K. Satoh, & C. H. Smith (Eds.), Discovery science: Proceedings of the 5th international conference (pp. 178-191). Berlin: Springer-Verlag.