UCSB Researchers Secure Funding for New AI Frameworks
The $1 million grant aims to move AI models beyond pairwise interactions to better map complex scientific laws.
Updated on Oct. 21, 2026 in Mathematics
Live Poll
Do you believe AI is currently making significant progress in solving complex scientific challenges?
Four researchers at the University of California, Santa Barbara have received a $1 million National Science Foundation grant to develop new AI frameworks. The project seeks to move beyond standard two-component modeling to better understand complex, multi-component scientific systems.
Why it matters
Current AI models are often constrained by their inability to analyze interactions beyond pairwise relationships, limiting their scientific utility. This research aims to create more robust models capable of identifying complex, higher-order scientific laws.
The $1 million grant supports the development of AI methods utilizing topological deep learning—a framework for analyzing the shape of data—and Kolmogorov-Arnold networks, which offer an alternative to traditional multi-layer perceptron neural networks.
The players
Nina Miolane
A researcher at the Robert Mehrabian College of Engineering focusing on the mathematical foundations of AI.
Haewon Jeong
A researcher at the Robert Mehrabian College of Engineering contributing to the development of higher-order scientific AI models.
Sanjukta Krishnagopal
A researcher at the Robert Mehrabian College of Engineering specializing in complex system interactions.
Yao Qin
A researcher at the Robert Mehrabian College of Engineering working on the development of robust AI frameworks for science.
National Science Foundation
The federal agency supporting fundamental research and education in all non-medical fields of science and engineering.
The details
The project addresses the limitation of current AI models that typically limit analysis to pairwise interactions between two variables. The researchers are building frameworks that account for overlapping, multi-component system dynamics. They will test these methods using synthetic and real-world datasets across neuroscience, molecular biology, and cosmology to ensure the models provide provable, reliable insights for scientific discovery.
Timeline
Earlier this year: The Real AI initiative launched at UC Santa Barbara.
October 21, 2026: The National Science Foundation grant award was reported.
The Tech Race
This project marks the first federal funding milestone for the recently launched Real AI initiative at UC Santa Barbara. It places the team in a competitive field seeking to bridge the gap between pattern-matching machine learning and the rigorous proof required for fundamental scientific laws.
The research team plans to release their developed tools as open source software once complete. This will eventually allow scientists and students to apply new AI frameworks to their own datasets in fields ranging from biology to cosmology.
The takeaway
The team's success in securing this grant signals a shift toward prioritizing mathematically provable robustness in scientific AI. Readers should watch for the eventual release of their open source framework, which will provide a new tool for modeling complex multi-component interactions.
Further reading
For more on how computational models are evolving, see Mathematics.
Source note: This article includes information reported by Edhat Santa Barbara.
Live Poll
Do you believe AI is currently making significant progress in solving complex scientific challenges?
