Nhat Ho Receives 2026 COPSS Emerging Leader Award

Nhat Ho of the Department of Statistics and Data Science has been named a 2026 recipient of the COPSS Emerging Leader Award. The Committee of Presidents of Statistical Societies presented the honor at the 2026 Joint Statistical Meetings in Boston. Ho was one of eight statisticians recognized this year.

The committee’s citation for Ho reads in full: “For pioneering the statistical foundations of modern AI, particularly convergence and identifiability of mixture-of-experts, advances in computational optimal transport, and theory for parameter-efficient fine-tuning of foundation models; for commitment to teaching, mentorship, professional services, and outreach.”

The award was established in 2020 for early-career statistical scientists who show evidence of, and potential for, leadership in the field. No more than eight people receive it each year. Notably, the award’s founding committee was chaired by SDS Professor Catherine Calder.

Research That Underpins Modern AI

A major theme of Ho’s recent work is understanding the statistical principles behind modern AI systems: when their internal structure can be identified, how efficiently they can be learned, and when their predictions and representations are stable. His research is guided by four principles:

• Data heterogeneity: Real-world data is complex and messy, and models need to handle that variety.
• Interpretability: Understanding why a model makes the predictions it does.
• Stability: Making sure results hold up rather than shifting with small changes.
• Scalability: Building methods that keep working as data and models grow.

Mixture-of-experts models divide a complex task among many specialized sub-models coordinated by a learned gating mechanism. They have become central to scaling up AI efficiently. Ho’s research establishes when the parts of such models can be reliably identified and how quickly they can be learned from data.

Ho’s work on computational optimal transport, a mathematical framework for comparing probability distributions, makes those methods faster and more practical at scale. His theory for parameter-efficient fine-tuning addresses how enormous pretrained models can be adapted to new tasks by updating only a small fraction of their parameters.

A Career Built Across Institutions

Ho earned his Ph.D. in statistics from the University of Michigan in 2017. He then worked as a postdoctoral fellow in the Department of Electrical Engineering and Computer Science at the University of California, Berkeley. He joined The University of Texas at Austin in 2020 as an assistant professor and has since been promoted to associate professor. 

Beyond the department, Ho is a core member of the Machine Learning Laboratory and senior personnel with the Institute for Foundations of Machine Learning, reflecting the collaborative reach of his work across campus. He is known at SDS and beyond as a committed mentor for students and early-career researchers.

Please join us in congratulating Professor Ho on this well-deserved honor. Learn more about SDS faculty and student accomplishments at https://stat.utexas.edu/news.