Designing Robust, Interpretable, and Fair Social and Public Health Interventions

In the last decades, significant advances have been made in AI, ML, and optimization. Recently, systems relying on these technologies are being transitioned to the field with the potential of having tremendous influences on people and society. With increase in the scale and diversity of deployment of algorithm-driven decisions in the open world come several challenges including the need for robustness, interpretability, and fairness which are confounded by issues of data scarcity and bias, tractability, ethical considerations, and problems of shared responsibility between humans and algorithms. In this talk, we focus on the problems of homelessness and public health in low resource and vulnerable communities and present research advances in AI, ML, and optimization to address one key cross-cutting question: how to allocate scarce intervention resources in these domains while accounting for the challenges of open world deployment? We will show concrete improvements over the state of the art in these domains based on real world data. We are convinced that, by pushing this line of research, AI, ML, and optimization can play a crucial role to help fight injustice and solve complex problems facing our society.