Neal studied physics, with a focus on computational physics and applied mathematics, during his undergraduate education (at the University of Notre Dame). He completed his Ph.D. at Yale University (with Prof. Julien Berro) in Molecular Biophysics & Biochemistry and Physical & Engineering Biology. For his Ph.D., he used quantitative microscopy, image analysis, and mathematical modeling to study molecular assembly during the cellular process of endocytosis.
Neal was formerly a Postdoctoral Fellow in Internal Medicine and Computer Science at Yale University. He worked with Professor David van Dijk to develop new deep learning (DL) methods, clinicians (Dr. Andrew Taylor, Dr. Tariq Ahmad, and Dr. Nihar Desai) to develop ML methods for interpretable predictions based on longitudinal EHR, registry, or clinical trial data, and biologists (Prof. Craig Wilen, Prof. Akiko Iwasaki, and Prof. Janghoo Lim) to identify molecular & cellular determinants of biological state using DL and single-cell omics data.
Neal has taught mathematical methods & programming, including statistics & classical machine learning, to graduate students and experimentalists and is eager to continue teaching in the Bay Area. Neal co-developed a novel self-supervised learning framework to generate de novo edge features from any graph and ingest them into an architecture that consistently achieves state-of-the-art performance on single-cell metadata label prediction tasks (paper). As follow up work, he is interested in developing GNNs for biomedical applications. Neal has also co-developed permutation invariant networks to gain insight into the properties of embedded spaces (paper), developed clinical decision support models (e.g., The COVID-19 Severity Index, which was incorporated into MDCalc), and ML methods to study the dynamics of single-cell transcriptomics data (e.g., Single-cell Longitudinal Analysis of SARS-CoV-2 Infection in Human Airway Epithelium). Neal hopes to build ML and data science methods for translational and biomedical research applications, perhaps by deconstructing and simplifying previous works, in order to add to an interdisciplinary research oeuvre.
- Patient phenotyping
- DL for single-cell- and multi-omics analysis
- Decolonial AI