I'm interested in societal implication of technology, especially machine learning. I'd like to help make automated systems more understandable and easier to use responsibly. I've also worked on applications of machine learning in healthcare.
- Harini Suresh, John Guttag.
A Framework For Understanding Sources of Unintended Consequences in Machine Learning. In Submission.
- Harini Suresh*, Jen Gong*, John Guttag.
Learning Tasks for Multitask Learning: Heterogenous Patient Populations in the ICU. Conference on Knowledge Discovery and Data Mining (KDD) 2018, London. ACM Conference Proceedings.
- Willie Boag, Harini Suresh, Leo Celi, Peter Szolovits and Marzyeh Ghassemi. Racial Disparities and Mistrust in End-of-Life Care. Machine Learning for Healthcare Conference 2018, Stanford CA. JMLR Workshop and Conference Track.
- Harini Suresh, Nathan Hunt, Alistair Johnson, Leo Anthony Celi, Peter Szolovits, Marzyeh Ghassemi. Clinical Event Prediction and Understanding using Neural Networks. Machine Learning for Healthcare Conference 2017, Boston MA. JMLR Workshop and Conference Track.
Workshops and Posters
- Willie Boag, Harini Suresh, Leo Celi, Peter Szolovits and Marzyeh Ghassemi. Modelling Mistrust in End-of-Life Care. Fairness, Accountability, and Transparency in Machine Learning Workshop, ICML 2018, Stockholm.
- Harini Suresh*, Divya Shanmugam*, John Guttag. Disparities in the Performance of Natural Language
Processing Tools. Women in Machine Learning (WiML) Workshop, NIPS 2017, Palm Springs CA.
- Harini Suresh, Peter Szolovits, Marzyeh Ghassemi. The Use of Autoencoders for Discovering Patient Phenotypes. Machine Learning for Healthcare Workshop, NIPS 2016, Barcelona.
- Harini Suresh. Feature Representations for Predicting ICU Mortality (SuperUROP thesis, Computational Biophysics Group, MIT RLE)
- ICU Intervention Prediction: MIT News, NVIDIA, Huffington Post
- Fair ML: MIT News