(Colloquium) Andrew Ackerman, University of North Carolina at Chapel Hill
November 15 @ 4:30 pm - 5:30 pm
Colloquium Talk
Andrew Ackerman, University of North Carolina at Chapel Hill
Friday November 15, 2024 at 4:30PM
Rocky 312
Title: Measures of Fairness and High Dimensional Data Integration
Abstract:The first component of this talk will present a representative discussion from a novel course, entitled Moral Machine Learning, developed at the University of North Carolina at Chapel Hill. In particular, we motivate and introduce statistical measures of fairness used to assess classification algorithms. This discussion will culminate in an Incompleteness Theorem, which demonstrates that these measures are, in some fundamental way, not totally reconcilable. How to assess fairness despite this incompleteness result will motivate open questions discussed at the conclusion of the second component of this talk. This latter component will primarily be focused on original research. We present completed work for high dimensional data integration for human neuroscience. In particular, neuroimaging studies, such as the Human Connectome Project (HCP), often collect multifaceted data to study the complex human brain. However, these data are often analyzed in a pairwise fashion, which can hinder our understanding of how different brain-related measures interact. In this study, we analyze the multi-block HCP data using the Data Integration via Analysis of Subspaces (DIVAS) method. We integrate structural and functional brain connectivity, substance use, cognition, and genetics in an exhaustive five-block analysis. This gives rise to the important finding that genetics is the single data modality most predictive of brain connectivity, outside of brain connectivity itself. Moreover, investigations of shared space loadings provide interpretable associations between particular brain regions and drivers of variability, such as alcohol consumption in the substance-use data block. Novel Jackstraw hypothesis tests are developed for the DIVAS framework to establish statistically significant loadings. We conclude by discussing proposed future work, at both the faculty and undergraduate levels, in each of data integration and algorithmic fairness.