Human-centric Machine Learning in Healthcare

Nina Corvelo Benz (Ph.D. Student)

With machine learning tools being widely used in social and commerce systems and increasingly also to automate processes in other sectors such as medicine, banking, and employment, machine-made decisions are becoming more and more influential to our everyday life. In many of these real-world applications algorithms’ decisions affect human behavior and vice versa leading to decision and data dependency loops. Current machine learning models however are often designed for full automation disregarding many aspects of the interaction between humans and systems. Consequently, the issues regarding fairness, robustness and transparency of these tools are especially evident in co-dependent applications. My research will concentrate on the topic of human-centric machine learning. The objective is to tackle before-mentioned issues by accounting for human factors such as feedback loops, strategic decision makers seeking for beneficial outcomes, and different levels of automation. For example, increasing robustness and trust by developing models that defer decisions with high uncertainty to humans or provide actionable insights about decisions taken without incentivizing adverse behavior. In particular, I am interested in the application of such models in healthcare and bioinformatics. It is a complex application domain for machine learning due to the heterogeneity and high dimensionality of the data. Hence, the limits of fully automated models become apparent in the face of small amounts of data sets and rare events.

Primary Host:	Manuel Gomez Rodriguez (Max Planck Institute for Software Systems)
Exchange Host:	Karsten Borgwardt (Max Planck Institute of Biochemistry)
PhD Duration:	01 June 2021 - 30 June 2025
Exchange Duration:	01 January 2023 - 30 June 2023 - Ongoing