May 27 (Wed) @ 1:00pm: "Towards Trustworthy Machine Learning: Fairness Under Dynamics and Counterfactuals," Ozgur Guldogan, ECE PhD Defense

Location: Engineering Science Bldg (ESB), Room 2001
Zoom: https://ucsb.zoom.us/j/7748747013

Abstract

Machine learning systems increasingly make decisions with long-term consequences for individuals and society. Designing trustworthy machine learning systems therefore requires going beyond static accuracy and one-shot fairness criteria. Decisions may alter future populations through feedback, individuals may improve and reapply, uncertainty may call for set-valued predictions rather than point estimates, and large-scale deployment may be constrained by the efficiency of modern generative models. This talk will focus on Counterfactually Fair Conformal Prediction, a framework for constructing prediction sets that are both statistically valid and counterfactually fair. By aggregating conformity scores across counterfactual interventions on protected attributes, the method produces prediction sets that remain invariant under counterfactual changes while preserving finite-sample marginal coverage guarantees. The talk will also briefly describe Multi-Bin Batching, a length-aware inference-serving policy that improves large language model throughput by grouping requests according to predicted generation lengths.

Bio

Ozgur Guldogan is a PhD candidate in the Department of Electrical and Computer Engineering at UC Santa Barbara, under the supervision of Prof. Ramtin Pedarsani. He received his B.S. degrees in Electrical and Electronics Engineering and Mathematics from Bogazici University in 2021, and his M.S. in Electrical and Computer Engineering from UC Santa Barbara in 2024. He is a recipient of the UC Regents Fellowship at UC Santa Barbara. His research interests include trustworthy machine learning, fairness in decision-making, and conformal prediction.

Hosted By: Professor Ramtin Pedarsani

Submitted By: Ozgur Guldogan <ozgurguldogan@ucsb.edu>