The University of Georgia’s Institute for Artificial Intelligence (IAI) recently announced the recipients of the latest round of interdisciplinary seed grants. College of Public Health Assistant Professor Chao Huang was awarded a grant for his project titled “AI-driven multimodal biomarker discovery and reasoning in depression and related mental health disorders.”
Huang’s research primarily focuses on statistical learning of large-scale biomedical data, including clinical, imaging and genomic data. He aims to novel statistical methods and machine learning algorithms for analyzing data with complex structures.
Epidemiology & Biostatistics
Assistant Professor Chao Huang.
This is the second year of the Seed Grant Program Awards. Refined by a careful review process, the competition considered innovative ideas from across various colleges at the university. The grants totaled $276,250 through financial support of the Office of the Senior Vice President for Academic Affairs and Provost, the Office of Research and the College of Agricultural and Environmental Sciences.
The honored projects align with the IAI’s five convergence themes, which are AI in Education, Ethics of AI for Human Society, AI and Future of Health & Work, AI for 3F, and AI for Cyber & Societal Security. Huang’s project was awarded under the “AI and Future of Health & Work” theme.
Co-authored with faculty from the School of Computing Research Professor Tianming Liu and Franklin College of Arts and Sciences Assistant Professor Rongjie Liu, this project envisions a future in which “explainable and scalable AI systems enable precision psychiatry through more accurate diagnosis, mechanistic insight, and data-driven clinical decision support.”
The project’s objective is to develop an AI-driven multimodal framework through the integration of neuroimaging and clinical, behavioral and genetic data that improves detection, characterization and mechanistic understanding of depression and related mental health disorders. It also aims to position UGA as a leader in AI-driven mental health research and interdisciplinary workforce development.
The project timeline is 6 months from Jan. 1, 2026 to June 30, 2026. It will approach its goals through first developing a multimodal Transformer model designed to integrate neuroimaging and clinical, behavioral and genetic data for depression risk and prediction. Secondly, a depression-specific large language model (LLM) that converts structured multimodal features into patient-level diagnostic summaries.
Model training, validation and testing took place in months one and two, aiming for a multimodal transformer for depression risk discovery and prediction. Model training, validation and testing is set to continue in months three and four, pursuing a depression-specific LLM for automated clinical report generation. Dissemination will take place in months four through six.
By Alexia Rule
