November 8 Colloquia: "On the implausibility of bias-preserving algorithms", a talk by Kino Zhao

Abstract: Discussions of algorithmic bias often assume, without reflection, that biases in predictive outcomes reflect biases in society. That is, machine learning algorithms are biased because they are bias-preserving processes. This talk challenges this assumption by pointing to a tension between the philosophy of data literature, which largely rejects the idea that data possess essential features that are always preserved during analysis, and the claim that social biases are easily preservable despite attempts at getting rid of them. I further discuss some benefits for understanding algorithmic bias as a data feature problem.

Speaker Bio: Kino Zhao is an Assistant Professor at Simon Fraser University Department of Philosophy. Her research is in the philosophy and methodology of the social sciences. She is interested in both philosophically-informed-science and scientifically-informed-philosophy. Her projects seek to either: 1) identify a methodological concern discussed by social scientists and see if providing a philosophical analysis can help the dialogue make progress, or 2) identify a philosophical thesis that seems to involve factual claims about the social world and ask whether the social sciences are equipped to supply them. She can also be persuaded to undertake/supervise projects in philosophical logic, philosophy about machine learning, game and decision theory, most areas of technical philosophy, and historical topics around the Vienna Circle.

Join online here.