Mathematics for a Better World: Richard Tsai Earns 2018 Peter O’Donnell Distinguished Researcher Award

Imagine a small UAV drone—the kind you see kids piloting at parks—whizzing through the halls of the New York Metropolitan Museum of Art.

There’s no behind-the-scenes operator, nor pre-programmed flight path controlling the drone as it makes its way past the priceless works. It’s machine learning and data analysis deployed in real-time that enables the vehicle to move through the halls, inspecting every inch of the museum—including the artwork—without risking a run-in with a Rembrandt in the process.

This is the scene ICES Professor Richard Tsai paints when describing how his machine learning algorithms could help devices like UAVs navigate new environments. In addition to art museums, he gives an example of the drones being deployed to inspect hard-to-reach infrastructure, such as the underside of bridges.

The point is for the technology to work in a wide variety of applications and environments, with Tsai’s mathematical research providing a foundation for the responsiveness and flexible programs that make it possible. That applies to the UAV research, as well as a whole host of other innovative research streams. Data analysis techniques could help physicians find and flag patient anomalies in colonoscopy images. It could develop more efficient methods for solving computations on parallel processors. It could create numerical algorithms that emulate behavior at interfaces (where one material ends and another begins) by developing computational models for crystal growth.

“I try not to follow, and I want to do something that’s useful,” said Tsai, who also holds an appointment in UT's Department of Mathematics. “As a mathematician, it’s possible to say that we can understand different aspects [of scientific problems].”

Tsai’s enthusiasm for exploring how applied mathematics can help explore subjects and solve problems across the sciences was recently recognized with the 2018 Peter O’Donnell Distinguished Research Award. The distinction recognizes sustained research contributions by an individual, while providing $100,000 over four years for future research.

“Professor Tsai is a valued member of ICES and the ICES Center for Numerical Analysis, whose work is central to much of the ICES research agenda,” said ICES Deputy Director Robert Moser. “His contributions to ICES and the university are extraordinary, and we are delighted to be able to recognize him with the 2018 Peter O'Donnell Distinguished Researcher Award."

Tsai said that he plans to use the research funds in part to bring more speakers to the university to present their work and provide more training opportunities for students.

“This kind of award really increases research activities,” said Tsai. “You can invite people for an extended amount of time, which is very valuable.”

Tsai’s research spans the entire process of computational problem-solving, from developing algorithms to model phenomena from processing the algorithms on high-performance computers. He said that machine learning—algorithms that respond in real-time to the data that’s being acquired, and in turn, makes adjustments to a program—are the “new toy” in computational science, and an area where he has been focusing much of his recent research. In fact, he recently submitted a research paper outlining how machine learning can benefit mathematics research as a whole.

The algorithms that he has developed so far involve creating programs that predict the potential value of new information plucked from a three-dimensional environment. Programmed into a UAV, or other autonomous vehicle, it could help the device sort out relevant information from background features in an environment, allowing it to zip through an array of an environments (perhaps even art museums).

Currently, this particular class of algorithms is still in the research phase, Tsai said. He’s working on enhancing the algorithms so they can handle environments that change over time and identify different obstacles and features. Nevertheless, the potential of mathematics to improve machine-learning applications has already brought the support of the National Science Foundation, the U.S. Army Research Office, and UAV companies.

Along with recognizing Tsai’s record of research excellence, The Peter O’Donnell Distinguished Research Award will help push research forward by providing funds that can help inspire and carry out new work.

“We’re often limited by the available resources,” Tsai said. “This kind of award helps us expand.”