Three computer science faculty win AI research awards

Chien-Ju Ho, Yevgeniy Vorobeychik and William Yeoh received $195,000 from J.P. Morgan Chase to support their work in artificial intelligence

Shawn Ballard 

Three faculty members in the Department of Computer Science & Engineering in the McKelvey School of Engineering at Washington University in St. Louis have been awarded J.P. Morgan AI Research Awards, given annually to leading researchers in artificial intelligence.

Chien-Ju Ho and Yevgeniy Vorobeychik won a faculty research award to study challenges in selecting a representative population with applications in medicine and machine learning. William Yeoh also won a faculty research award to improve user experience through goal recognition and explainable assistance in adaptive systems using artificial intelligence. 

In sponsoring the awards, J.P. Morgan Chase (JPMC) aims to advance cutting-edge AI research to solve real-world problems through ongoing partnerships in research and innovation between university faculty and JPMC.

Ho and Vorobeychik to tackle representative selection problem

Chien-Ju Ho, assistant professor, and Yevgeniy Vorobeychik, associate professor, received a $95,000 grant from JPMC to support their new study on the representative selection problem. Selecting a representative population is an essential step in conducting high-quality research in a range of scientific fields.

“Inequitable representation is one of the root causes of many biased and unfair results in scientific development,” Ho said. “This award enables us to develop approaches to mitigate the problem and could be impactful both in the medical domain, for example, by recruiting representative populations for medical trials, and in machine learning, for example, through curating representative datasets to train and evaluate machine learning models.”

Representative population selection helps ensure that results are generalizable, equitable and nondiscriminatory. In this project, Ho and Vorobeychik will use their expertise in behavior modeling and optimization to address key challenges in representative selection: the combinatorial nature of the problem and the uncertainty associated with the selection. Because there are many factors that play into representation, such as race, age and gender, and the outcome of a particular selection can’t be known in advance, the team will explore both the theoretical properties and the empirical performance of their proposed solutions under a range of practically-motivated settings. 

Yeoh to advance AI-assistive technologies

William Yeoh, associate professor, received a $98,924 grant from JPMC to develop AI technologies to assist humans in navigating complex systems. Yeoh’s new project will focus on goal recognition and explainability technologies to recognize a user’s goals based on their behavior, provide structured assistance to help the client reach their goals more effectively, and clearly explain why the assistance was provided. 

“I am very excited to collaborate with fintech experts at JPMC as they can provide the domain expertise and use cases necessary for transitioning and deploying the algorithms developed in the real world,” Yeoh said. “If findings from this project can help simplify the process of navigating JPMC's websites, apps or other systems, I would consider that a win, especially since there are millions of households who are JPMC customers.”

This new project draws on Yeoh’s prior experience in goal recognition, explainable planning, model reconciliation and logical inference methods. More broadly, Yeoh’s research program focuses on human-AI collaboration, for which improved user experience is a key requirement.

Click on the topics below for more stories in those areas

Back to News