While AI is assuming omnipresence today more than ever, its adoption is still limited in socially critical applications such as in public health, especially among low-resource and underserved communities. This talk presents how techniques from reasoning, strategic decision-making, or planning in uncertain, stochastic, or resource-limited settings, can be leveraged for tackling two such real-world public health challenges: tuberculosis prevention and improving maternal and child healthcare.  

Towards achieving this real-world impact, several fundamental research questions must first be addressed. For instance, community health workers and NGOs operating with limited health resources face the challenge of optimally utilizing these resources to maximize their impact. In doing so, such NGOs must account for domain-specific considerations such as fairness or risk-averseness and plan the limited resources to serve beneficiaries at scale, in an uncertain and dynamically changing world. In addition, accurate evaluation of novel solutions being built through Randomized Controlled Trials (RCTs) remains difficult due to high sample variance in these settings.  

Towards tackling these challenges, this talk presents how techniques such as Restless Multi-Armed Bandits (RMAB) can be utilized to optimize allocation of scarce health intervention resources. Transcending traditional research boundaries, this talk presents how this work has been transitioned from the blackboard to a first-of-its-kind field evaluation of the RMAB algorithm, involving 23,000 real-world mothers over a 7-week period, results of which show a ∼ 30% improvement in the performance metric of interest.