An Effective Theory of Bias Amplification

Описание: Date Presented: 9/16/25
Speaker: Arjun Subramonian, Meta FAIR

Visit links below to subscribe and for details on upcoming seminars:
https://www.isi.edu/isi-seminar-series
https://www.isi.edu/events

Abstract: Machine learning models can capture and amplify biases present in data, leading to disparate test performance across social groups. To better understand, evaluate, and mitigate these biases, a deeper theoretical understanding of how model design choices and data distribution properties contribute to bias is needed. In this talk, I will discuss how we developed a precise analytical theory in the context of ridge regression, both with and without random projections, where the former models feedforward neural networks in a simplified regime. Our theory offers a unified and rigorous explanation of machine learning bias, providing insights into phenomena such as bias amplification and minority-group bias in various feature and parameter regimes.

Speaker's Bio: Arjun Subramonian is a Research Scientist at Meta FAIR. They currently study the measurement validity of AI evaluations. In the past, their research has focused on the fairness and ethics of machine learning and natural language processing. They are a recipient of a FAccT 2023 Best Paper Award.