AI’s Inflection Point — The Texas Scientist

By Marc Airhart.

There’s a moment when everything changes, like when you’re rolling down the runway in an airplane. At first it just feels like being in a fast car, but then, at a certain velocity, the flaps shift, the wheels lose their grip on the ground and you feel buoyant. You’re soaring.

A similar liftoff moment has arrived for artificial intelligence at The University of Texas at Austin. After years of rising prominence, the university’s AI research and educational offerings have soared into new airspace, that rarefied realm known as preeminence.

A Growing Community

“In five years, the university has roughly doubled the number of artificial intelligence experts on the faculty.”

In five years, the university has roughly doubled the number of AI experts on faculty, spanning 77 independent labs across campus. And more are on the way. In the College of Natural Sciences alone, eight AI faculty have joined in recent years, boosting the total in the Department of Computer Science to 22.

In 2020, following an intense nationwide competition, the National Science Foundation selected UT Austin to lead the Institute for Foundations of Machine Learning. Machine learning is the technology that drives AI systems, enabling them to acquire knowledge and make predictions in complex environments.

“Machine learning has seeped into practically every area of science and industry, from video processing to drug design to finance,” said Adam Klivans, a professor of computer science who directs the new institute. “A key mission will be to develop core technologies that can break through obstacles common to all of these areas.”

The permanent base for campus-wide machine learning research, including the new NSF-supported institute, is UT Austin’s burgeoning new Machine Learning Laboratory. It brings together computer and data scientists, mathematicians, roboticists, engineers and ethicists to meet the institute’s research goals while also working collaboratively on other interdisciplinary projects.

Space to Play

One of the hottest application areas for AI right now is robotics. Unlike theorists, roboticists need space for their creations to explore.

On a sunny day last October, officials from the Army Futures Command were among those to tour the newly renovated Anna Hiss Gym, which is something like a robotics tinkering shop out of a sci-fi movie. The new space is home for Texas Robotics, a campus-wide research and education consortium. Its experts partner with industry groups as well as with the Austin-based Army Futures Command, the new four-star command that is transforming Army modernization through thoughtful collaborations with researchers, including the UT Austin robotics experts.

“My vision for this space is for it to be a living laboratory and showcase for all of the robotics research on campus, a space that fosters interdisciplinary cohesion and world-class creative research and also serves as a destination for the public to learn about robotics,” said Peter Stone, a professor of computer science and director of Texas Robotics.

Stone, whose own research has involved soccer-playing robots and self-driving cars, has seen astounding growth in robotics at the university in the last six years, since six core faculty members first banded together to launch a graduate portfolio program and coordinate research and hiring.

“Since then, we’ve launched a successful industrial affiliates program, partnered in a huge and ambitious way with Army Futures Command and grown to 16 core faculty who collectively cover all areas of robotics, with particular emphasis on what we see as the most important growth areas for the field: long-term autonomy and human-robot interaction,” Stone said.

Making it Good

As AI becomes more and more a part of our daily lives, a wide-ranging team from across the university – computer scientists, information architects, philosophers, transportation experts, communicators, ethicists, policy researchers and others – are trying to hash out guiding principles for building systems likely to have a positive impact and fewer unintended consequences. The Good Systems initiative was launched in 2019 as one of the university’s “Bridging Barriers” grand challenges.

One project seeks to teach robots how to learn new tasks by imitating humans, and to do it safely and efficiently.

“Nobody is going to want to put a robot in the home if they aren’t safe,” said Scott Niekum, assistant professor of computer science.

Multiple projects tackle the problem of biases in data that can lead to AI systems that perpetuate human biases when making critical decisions in hiring, for example, or in criminal justice. Another project is designing an AI system that optimizes the routing of ambulances during disasters, such as the COVID-19 pandemic, to save lives.

Training the Next Generation

U.S. News and World Report has ranked UT Austin fifth in the country for undergraduate AI programs and in the top eight for graduate programs. In a state known for superlatives, the university stands out as the only one in Texas ranked in the top tier for undergraduate or graduate programs in AI. UT’s Artificial Intelligence Laboratory, led by professor Ray Mooney, has been a driving force behind that distinction by recruiting many of the top names in the field.

“It is quite powerful to have leading researchers doing formal instruction in the classroom at not only the graduate level, but also the undergraduate level,” said Kristen Grauman, a professor of computer science and one of three fellows of the American Association for Artificial Intelligence on faculty. “It makes the subject come alive for our students when they can learn and use cutting-edge methods in their labs and assignments. This is really important in a rapidly evolving field like AI.”

Students also benefit from state-of-the-art facilities, including the supercomputing resources of the Texas Advanced Computer Center, home to the world’s most powerful academic supercomputer, and the expertise housed in the Oden Institute for Computational Engineering and Sciences.

With a passion to change the world, artificial intelligence researchers at UT – world-class scientists and students together – aim for the sky and reach altitudes most can only dream of.

Take a video tour through a Texas Robotics space with UT Austin graduate students.

“It makes the subject come alive for our students when they can learn and use cutting-edge methods in their labs and assignments.”