By Rachel Robey
Breakthroughs happen when researchers collaborate and share data, but it’s hard to convince different organizations to trust each other. Kandasamy addresses this challenge — and society reaps the benefits.
The Department of Computer Sciences (CS) is pleased to announce that Assistant Professor Kirthevasan Kandasamy is the recipient of a 2025 National Science Foundation (NSF) CAREER Award for his proposal “Game-theoretic Foundations for Incentive-aware Data Sharing and Collaborative Machine Learning.” NSF’s CAREER Award is one of the most prestigious honors early-career faculty can receive, and is reserved for “academic role models” whose potential includes — and transcends — excellence in education and research.
“Federal support is crucial to work like this — work that has no immediate commercial payoff. Industry may not be incentivized to study these issues in order to maintain their competitive edge,” says Kandasamy, who joined the CS faculty in 2022. “However, the long-term impact of work like this could benefit all organizations — and, ultimately, society as a whole — by creating a foundation for more open, cooperative, and socially responsible data use.”
Building brain trust
A theoretical researcher in machine learning and game theory, Kandasamy investigates the incentives and dynamics at play when organizations share data with one another. His current work, supported in part by the CAREER Award, develops mechanisms that promote truthful data contribution, ensure fair benefits for the participants, and generate value for society as a whole.
“Today, many breakthroughs in areas like health care, education, and public policy depend on combining data from different sources — but that only works if everyone involved feels confident that the collaboration is fair, worthwhile, and that others are acting in good faith,” he explains.
For example: Suppose you represent a pharmaceutical company with troves of data on a particular medical condition. Your competitors also study this condition and have their own caches of data. As competitors, sharing this data would be unusual, and the lack of trust between you would make it all the more unlikely. But imagine if you did share this data – the medical breakthroughs and commercial value of potential solutions could be unprecedented.
This is where Kandasamy comes in: The goal of his research is to build the trust necessary for these collaborations to take place.
Data sharing as a game
Through a blend of game theory and machine learning, Kandasamy identifies the conditions under which organizations might be willing to share data. From there, he creates mechanisms that ensure the resulting data sharing is honest (i.e., by preventing people from supplying corrupt, misleading, or fabricated data) and equitable (i.e., the amount of data you receive is in direct proportion to the amount you supply). By imagining data sharing as a game, Kandasamy establishes rules of play that incentivize honest participation.
“In the long run, this could help scientists and organizations collaborate more effectively, leading to discoveries that none of them could have achieved alone,” he says.
Theory meets practice in Kandasamy’s current collaboration with researchers in the Department of Biostatistics and Medical Informatics (BMI) and the Alzheimer’s Disease Research Center (ADRC). The ARDC is part of a research consortium that includes other universities and hospitals, and while all share a long-term goal of treating — even beating — the disease, the individual researchers and organizations are naturally protective of their own data and results. Kandasamy supports data sharing initiatives across the consortium, enabling democratized data by designing machine learning protocols that determine whether data is complete, accurate, honest, fair, and valid — in a word, “good.”
“The techniques we use are based on machine learning, because we’re collecting all this data and trying to reason about its quality,” he explains. “From a statistical perspective, it’s very challenging to validate data in this way.”
Congratulations again — and best of luck — to Professor Kandasamy.
About Kirthevasan Kandasamy: Kirthevasan Kandasamy is an assistant professor in the Department of Computer Sciences at the University of Wisconsin–Madison. His research interests are in the intersection of machine learning and game theory. He was a postdoctoral scholar at the RISE Lab at the University of California, Berkeley working with Ion Stoica, Mike Jordan, and Joey Gonzalez. He completed his PhD in Machine Learning at Carnegie Mellon University, where he was co-advised by Jeff Schneider and Barnabas Poczos. His PhD was supported by a Facebook fellowship (2017/18), a Siebel scholarship (2017/18), and a CMU Presidential fellowship (2015/16). Prior to CMU, he completed his B.Sc in Electronics & Telecommunications Engineering at the University of Moratuwa, Sri Lanka.
About the NSF CAREER Award: The Faculty Early Career Development (CAREER) Program is a Foundation-wide activity that offers the National Science Foundation’s most prestigious awards in support of early-career faculty who have the potential to serve as academic role models in research and education and to lead advances in the mission of their department or organization. Activities pursued by early-career faculty should build a firm foundation for a lifetime of leadership in integrating education and research.