The primary focus of my research is unsupervised learning based on latent factor discovery. The main application for this work is understanding high-dimensional but under-sampled data coming from human biology and behavior. 

• An information-theoretic foundation for modularly and hierarchically decomposing information in complex systems with prototype and sample applications (NIPS-14) and (code), and more theoretical developments (AISTATS-15). The linear version of CorEx exhibits a unique "blessing of dimensionality" for recovering latent factor structure and excellent performance for estimating covariance matrices with high-dimensional, under-sampled data (code). An incremental version, the information sieve, is introduced in (ICML-16) and (IJCAI-17).
• Some recent works explore connections between our information-theoretic learning ideas and error-correcting codes, invariant/fair representation learning (NeurIPS-18), semi-supervised topic models (TACL-17), and VAEs.
• We also study foundational issues about information theory and ML. (1) A preoccupation with (pairwise) mutual information leads to its frequent mis-use, see (ICML-14) for one example. (2) Information measures are hard to estimate, (UAI-15) (AISTATS-15) (NIPS-16).
• Applications: gene expression (interesting podcast and article about this work), neuroscience 1, 2, 3, 4, text analysis (code) 1 2, psychometrics, finance

See our group page for other ongoing efforts, or this page for more information about CorEx. I update my blog and twitter slightly more often than this page.