Computer Science Speakers Calendar

View Full Calendar

Integrating Human Knowledge in Machine Learning Models and Explanations

Event Type
The Department of Computer Science Machine Learning Seminar
1214 Siebel Center
Apr 5, 2024   2:00 pm  
Han Zhao


With the rapid increase in computational power in recent years, deep learning has achieved unparalleled success across various domains. However, deep models are often criticized for their black-box nature, particularly in critical applications where understanding the decision-making process is crucial. The talk will focus on integrating human knowledge through effective regularization of machine learning model behavior to ensure predicted outcomes are both understandable and reliable for humans. Firstly, I will introduce a novel method aimed at enforcing transformation equivalence in model explanations. Secondly, I will present a theoretical framework to improve distributional robustness, which encourages that model predictions rely on invariant features rather than spurious ones. Lastly, I will delve into the analysis of irreversible trajectories in longitudinal data, with applications in Alzheimer's disease studies. These techniques will enhance trust in AI through explainable, fair, and robust models. 


Xiaoqian Wang is an Assistant Professor of Electrical and Computer Engineering at Purdue University. She received her Ph.D. degree from the University of Pittsburgh in 2019, and the B.S. degree from Zhejiang University in 2013. She focuses on designing novel machine learning models for interpretability, fairness, and robustness. She also work on the intersection of machine learning and bioinformatics, healthcare. She received an NSF CAREER award in 2022, and an AAAI distinguished paper award in 2023. 

link for robots only