The following is a list of preprints and peer reviewed papers on applications of my research (machine learning, probability theory, algorithms) to problems in medicine - endotype identification, drug repurposing etc.
A resume is here.
List of all publications (ORCID): 0000-0002-2739-1246.
Applications of Machine Learning to Medicine
Note that author order is alphabetical (by last name) in the papers; and venues like ICML and Neurips are conferences, but they are rigorously peer-reviewed.
Some slides
-
Identifying endotypes for ulcerative colitis (slides)
[Summary: This is an application of my paper on a MAX-CUT based algorithm to identify endotypes in heterogenous diseases from sequencing data. It correctly identifies the existing OSM high (West et al. 2017) endotype and proposes novel ones as well.].
-
Identifying endotypes for chronic kidney disease.(slides)
[Summary: similar to ulcerative colitis above but for inflammatory kidney diseases and uses proteomics data as well. Adding proteomics allows downstream verification of the identified endotype.].
Papers/preprints on applications of neural networks to medicine:
-
Learning rate under distrbution shift (with an application to flow cytometry). Joint with M.Fahrbach, A.Javanmard, V.Mirrokni. ICML, 2023. (long version,
poster)
[Summary: We show that optimizing the learning rate for continuously trained neural networks can lead to better accuracy. We provide an application of our algorithm to flow cytometry using single cell RNA-seq data from murine pancreas (Bergen et al, 2020)].
-
Enhancing selectivity using Wasserstein distance based reweighing. Under submission, 2026.(preprint)
[Summary: We show that reweighting neural network training data can lead to much better chances of discovering selective drug molecules. In particular, we use our algorithm to predict small molecules which bind to MNK2 but not MNK1. Our predictions have a success rate of 5% (as opposed to less than 1% for non-ML methods) after synthesis and single point of concentration testing.]
Papers/preprints on applications of classical machine learning techniques for genomics and drug repurposing:
-
Designing optimal tests for slow converging Markov chains. Joint with C.Stein. IMLH (ICML), 2023. (short version)
[Summary: We design a hypothesis testing algorithm that can work with very small samples of data - as small as 20 cells. As a prototype example, using publicly available single cell RNA seq data (Bergen et al., 2020), we show that our algorithm can be much more stable than the usual hypothesis testing algorithm.]
-
Approximating regulatory pathways from non-sequential data for design of basket trials, joint with C.Stein, under submission, 2026. (preprint)
[Summary: We design an algorithm to compute shared biomarkers for drug repurposing via basket trials.]