I am a PhD candidate in the ECE department at Northeastern University and part of Dr. Jennifer Dy’s Machine Learning Lab at the SPIRAL research center. I’m broadly interested in improving transparency in black-box prediction models, specifically in relation to interpretability and uncertainty quantification.
I collaborate with the Channing Division of Network Medicine in applying machine learning methods to various challenges related to Chronic Obstructive Pulmonary Disease (COPD). COPD is a lung disease commonly associated with smoking and long-term exposure to lung irritants. We work with multi-omics, spirometry data, and lung imaging to improve subtyping, disease progression, diagnosis, etc.
- Black-Box interpretability, especially feature attribution and feature selection methods.
- Uncertainty Quantification in Neural Networks
- Variational Inference and Normalizing Flows
- Geometric Deep Learning
Explanations of Black-Box Models based on Directional Feature Interactions
Aria Masoomi, Davin Hill, Zhonghui Xu, Craig P. Hersh, Edwin K. Silverman, Peter J. Castaldi, Stratis Ioannidis & Jennifer Dy
ICLR 2022 (Spotlight, ~5% of submissions)
[Paper] [Poster] [Code]
Deep Learning Utilizing Discarded Spirometry Data to Improve Lung Function and Mortality Prediction in the UK Biobank
Davin Hill, Max Torop, Aria Masoomi, Peter J. Castaldi, Michael H. Cho, Jennifer Dy & Brian D. Hobbs
ATS 2022 (Oral, ~5% of submissions)
Explanation Uncertainty with Decision Boundary Awareness
Davin Hill, Aria Masoomi, Sandesh Ghimire, Max Torop & Jennifer Dy
Analyzing the Effects of Classifier Lipschitzness on Explainers
Zulqarnain Khan, Aria Masoomi, Davin Hill & Jennifer Dy
Leveraging deep-learning on raw spirograms to improve genetic understanding and risk scoring of COPD despite noisy labels
Justin Cosentino, Babak Behsaz, Babak Alipanahi, Zachary R McCaw, Davin Hill, Tae-Hwi Schwantes-An, Dongbing Lai, Andrew Carroll, Brian D Hobbs, Michael H. Cho, Cory Y. McLean & Farhad Hormozdiari