Frontiers in Machine Learning: Beyond Fairness: Pushing ML Frontiers for Social Equity [Panel]

At its core, machine learning is the artful science of statistically divining patterns from stores of data—typically, lots of data. Much of these data are drawn from sources as diverse as tweets and Creative Commons images to COVID-19 patient health records. Machine learning uses innovative techniques to draw what it can from the data on hand to push the boundaries of such problems as reliability and robustness in algorithmic modeling; theories and applications of causal inference; development of stable, predictive models from sparse data; uses of interpretable machine learning for course-correcting models that confound reason; and finding new ways to use noisy or sparse annotated training data to drive insights. While societal impact and social equity are relevant to the frontiers above, this panel asks: How might ML take up data and questions across a variety of domains such as education, development, discrimination, housing, health disparities, inequality in labor markets, to advance our understanding of systemic inequities and challenges? These systems, arguably, tacitly shape the data, theory, and methods core to ML. How might centering questions of social equity advance the frontiers of the field?

Moderator: Mary Gray, Microsoft

Speakers: Rediet Abebe, University of California, Berkeley
Irene Lo, Stanford University
Augustin Chaintreau, Columbia University

日期：

2020年7月21日

演讲者：

Mary Gray, Rediet Abebe, Irene Lo, Augustin Chaintreau

所属机构：

Microsoft Research, University of California Berkeley, Stanford University, Columbia University