Personalizing Population Health Through Machine Learning

The role of locality is critical in formulating effective public health responses; yet, the availability of localized public health data and predictions is frequently limited. Predominantly, decisions in public health are taken at state or national levels, neglecting the differences in needs between cities and their varying demographics. Stakeholders exist at each level, however, the pertinent information to guide their decisions is often presented at coarse granularities, resulting in limited local insight. Technological advancements enable us to collect granular data from novel sources. However, current methodologies lack mechanisms to integrate such diverse data. Moreover, establishing coherent connections between different scales and understanding their interrelations poses significant challenges.

My research envisions developing novel ML techniques to facilitate the integration of granular non-traditional datasets and enable more localized estimations and predictions. During this talk, I will discuss some of the methods we have developed and will engage the audience in thinking about the missing pieces of this vision and what advancements are required to realize its potential. Furthermore, I will discuss how AI can mitigate biases in public health data. This can help us further personalize public health interventions and tailor them to the diverse needs of various communities.

Zoom information will be sent to e-HAIL members.