ABSTRACT

Across several fields in science and engineering, we are increasingly turning to machine learning solutions for making decisions that can affect our lives in profound ways. Unlike many of these success stories, machine learning has had limited success in healthcare. Yet, healthcare is more important now than ever. In this talk, I will discuss the importance of evidence in building models for actionable healthcare. This evidence needs to i) account for various types of uncertainties and ii) be interpretable to a human decision-maker. First, I will present a framework for learning-to-defer pre-emptively to an expert in sequential decision-making settings that measures the likelihood of improving patient outcomes by deferring earlier versus deferring later, based on the underlying uncertainty in dynamics. On several healthcare applications, I will show SLTD outperforms existing non-sequential learning-to-defer baselines, whilst reducing overall uncertainty. Second, I will show how building small, inspectable models that humans can validate, can help us manage hypotension in the ICU. Together these tools may be seen as ways of obtaining evidence that can be used to vet reasonable hypotheses for actionable healthcare.

BIOGRAPHY

Dr. Parbhoo is an Assistant Professor and leader of the AI for Actionable Impact Group at Imperial College London, School of Electrical Engineering. Her research focuses on decision-making in uncertainty, causal inference and building interpretable models to improve clinical care and deepen our understanding of human health, with applications in areas such as HIV and critical care. Prior to moving to London, Sonali was a postdoctoral research fellow at Harvard, working with Prof Finale Doshi-Velez. Her work has been published at a number of machine learning conferences (NeurIPS, AAAI, ICML, AISTATS) and medical journals (Nature Medicine, Nature Communications, AMIA, PLoS One, JAIDS). She was also a Swiss National Science Fellow and was named a Rising Star in AI in 2021. Sonali received her PhD (summa cum laude) in 2019 from the University of Basel, Switzerland, where she built intelligent models for understanding the interplay between host and virus in the fight against HIV. Prior to this, Sonali completed her undergraduate and Masters at Wits University, South Africa where she specialized in Molecular Biology, Computer Science and Mathematics. Apart from her research, Sonali is also passionate about encouraging more discussion about the role of ethics in developing machine learning technologies to improve society.