Measuring Our Chances: Risk Prediction in This World and its Betters

Watch a thought-provoking lecture exploring the philosophical and mathematical challenges of algorithmic risk prediction, delivered by Harvard's Gordon McKay Professor Cynthia Dwork. Delve into the complex interpretation of probability in non-repeatable events and examine how AI algorithms generate life-altering predictions without clear mathematical foundations. Learn how recent research applying pseudo-random number theory offers potential solutions for evaluating prediction algorithms, while considering the broader implications of fairness in algorithmic decision-making. Explore the ambitious goal of approximating probabilities in an idealized, equitable world using only data from our systematically inequitable reality, and understand how this connects to fundamental questions about nondeterminism. Gain insights into how theoretical computer science can be applied to address critical societal challenges through this comprehensive 47-minute presentation from the 2023 Ding Shum Lecture series at Harvard CMSA.