Event Details

AI is now being used in many high-stake decision-making domains in which fairness is an important concern. In this talk, I will give a brief introduction on the issue of AI biases, where they come from, and the general strategies to measure and mitigate them. Specifically, I will talk about our work on the open-source AI Fairness 360 toolkit at IBM, as well as two bias mitigation algorithms that I co-developed. Lastly, I will also discuss our empirical work on explaining model decisions to make people more aware of AI biases.

Speaker Bio

Yunfeng Zhang is a Research Staff Member in IBM Research AI. His research interests lie in the intersection between HCI and AI. His recent projects focus on creating novel and effective AI explanation and bias mitigation algorithms, as well as investigating information designs that help people make better use of AI for decision tasks. He contributed to the development of both IBM’s AI Fairness 360 and AI Explainability 360 open-source toolkits, which are designed to help AI developers create intelligible and fair AI solutions. He received his Ph.D. in computer and information science from the University of Oregon in 2015.