I presented this work along with my teammate Armin Soltan as a poster at the AIM-AHEAD annual meeting, as well as an oral during the general session. We placed third overall and won $750! The competition was to predict breast cancer stage from whole-slide histology images in a completely held-out dataset balanced across racial groups.
More information about the NIH AIM-AHEAD Health Equity Data Challenge can be found here. Slides can be found here. Full-size poster can be found here.