Seminar by Dr. Wei Ding

When: 12:00 PM, Oct 7, 2026 (Wed)

Where: Web Lab

This presentation presents an executive-level view of how machine learning can transform complex health data into more meaningful and reliable health outcome assessment. Using physical activity analysis as a representative case, I will highlight our work on variable-length time-series summarization for wearable data and on fairness-aware deep learning that reduces performance disparities across individuals while preserving overall model quality. Together, these studies illustrate a broader vision for AI in health: moving beyond prediction accuracy toward systems that are robust, fair, interpretable, and better aligned with real-world healthcare needs.