UT San Antonio Researchers Develop Model to Improve Machine Learning Predictions for Dementia Across Diverse Populations
Scientists at The University of Texas Health Science Center at San Antonio have developed a new mathematical equation designed to enhance the generalizability of machine learning predictions, particularly in medical and neuroimaging research. This innovation addresses a significant challenge in applying AI models across diverse populations, especially in detecting conditions like dementia. The research, published in Nature Communications, highlights that traditional AI models often struggle with varied populations due to imbalanced data, leading to less reliable predictions. The new model proportionally weighs information, allowing for more accurate and generalizable predictions even when data points for certain groups are limited. The study's principal investigator, Mohamad Habes, PhD, emphasized the need for robust and objective AI models, particularly when data is scarce for specific demographics. Ngoc-Huynh Ho, PhD, the study's first author, created the mathematical model to guide the AI in learning pr...
