On Machine Learning for Prediction and Prioritization in the Allocation of Scarce Societal Resources

Explore how machine learning prediction methods can be ethically applied to allocate scarce societal resources in this 35-minute conference talk by Sanmay Das from Virginia Tech. Learn about the intersection of local justice theory and street-level bureaucracy with ML applications, discovering how separating measurement aspects from allocation strategies helps achieve societal goals. Examine real-world case studies from research on allocating services to households experiencing homelessness, providing K-12 educational supports, and managing caseworker time for eviction prevention support. Understand the critical differences in goals across various domains and gain insights into responsible implementation of predictive algorithms in social systems where resource scarcity demands careful prioritization decisions.