CMI Seminar: Nisheeth Vishnoi
TOWARDS CONTROLLING BIAS IN AI SYSTEMS
Powerful AI systems, which are driven by machine learning tools, are increasingly controlling various aspects of modern society: from social interactions (e.g., Facebook, Twitter, Google, YouTube), economics (e.g., Uber, Airbnb, Banking), learning (e.g., Wikipedia, MOOCs), to governance (Judgements, Policing, Voting). These systems have a tremendous potential to change our lives for the better, but, via the ability to mimic and nudge human behavior, they also have the potential to be discriminatory, reinforce societal prejudices, and polarize opinions. Indeed, recent studies have demonstrated that these systems can be quite brittle and generally lack the required qualities to be deployed in various human-centric/societal contexts. The reason being that considerations such as fairness, explainability, accountability etc. have largely been an afterthought in the development of such AI systems.
In this talk, I will outline our efforts towards incorporating some of the above-mentioned issues in a principled manner for core machine learning tasks such as classification, data summarization, ranking, personalization, and online advertisement. Our work leads to new algorithms that have the ability to control and alleviate bias from their outputs, comes with provable guarantees, and often has low "price of fairness".
Based on several joint works with Elisa Celis.