arXiv paper examining demographic bias in apparent age estimation models used for personalization. Compares MVL and AMRL distribution learning techniques across three datasets; AMRL achieves state-of-the-art accuracy but reveals persistent fairness-accuracy trade-offs, with significant performance degradation for Asian and African American populations.
Research
Apparent Age Estimation: Challenges and Outcomes
AMRL age estimation models achieve state-of-the-art accuracy while masking persistent demographic bias, with significant performance degradation for Asian and African American populations.
Tuesday, April 7, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.LG (Machine Learning)BY sys://pipeline
Tags
research