Abstract

In today's digital landscape, algorithms significantly influence our day-to-day decisions. However, as they gain prominence, their governance becomes essential to ensure fairness, reliability, and bias mitigation, especially in practical applications. Recognizing its importance, entities like China's nine ministries, the USA’s Office of Science and Technology Policy, and the European Parliament have already voiced their concerns.

My research delves into algorithmic governance in two distinct areas: (1) algorithm governance for predictive models from biased data and (2) algorithm governance for decision-making models. Predictive models are foundational to many systems, ranging from self-driving cars to online platforms. Biases in their training data can often result in undesired outcomes. I have formulated strategies to counter such biases, optimizing fairness and boosting model generalization. Beyond this, I also explore algorithms within decision-making facets of online systems—crucial in applications like online pricing and recommendation systems. My research touches on personalized pricing effects, shaping consumer protection policies, and analyzing information cocoons in recommendation algorithms. Through my talk, I intend to shed light on these insights, aiming for a future where algorithms are not just efficient but also trustworthy and equitable.

Time

2023-11-23  10:30 - 11:30   

Speaker

Renzhe Xu, Tsinghua University

Room

Room 308