Computer Science Researcher Receives NSF Grant

Lars Kotthoff has begun a project that aims to make developing artificial intelligence systems easier for researchers.

Kotthoff, an assistant professor in computer science at the University of Wyoming, will serve as the principal investigator for a National Science Foundation (NSF) grant project to explore algorithms and how they relate to artificial intelligence (AI) applications. The $412,000 grant will be in place until 2021.

Because algorithms are used in so many places in modern society, they are a key component to the nation’s economy. The applications include streamlining delivery of goods and language translation, but the scope of these become larger and more challenging every day. Kotthoff and the research team hope to enable developments in the area of algorithm selection, which involves automatically matching synergistic algorithmic choices to the specific properties of a problem to achieve optimal performance.

“There are often different approaches for solving the same type of problem, and they are often synergistic: where one fails, another performs well,” he writes in his abstract.

Current methods for making algorithm choices are limited. The issues are caused by reliance on brittle performance measures, limiting practical application in academia and industry. This project will address the limitations in three ways. First, it will define a notion of robustness to guide algorithm selection, and identify properties of algorithms, experimental setups, and computational environments that affect robustness. Second, it will develop specific performance measures informed by this definition of robustness, and which are portable across different hardware platforms. Third, it will mitigate the impact of brittle performance measures through new approaches to building performance models based on machine learning.

The project will result in the dissemination of shared data and benchmarks to the broader AI community.

“AI techniques in this project allow the best approach for a given problem to be chosen automatically,” Kotthoff writes. “This research will allow for such choices to be made more robustly even in difficult circumstances, resulting in improved performance and reduced effort to deploy AI in practical systems. Ultimately, the project will make it easier for humans to develop high-performance AI systems.”

Academic Summer School
Kotthoff had a busy summer, as he organized an academic summer school in Jackson, Wyo., on June 4-8.

The curriculum was focused on constraint programming, a sub-area of artificial intelligence, and was the 14th event in the series of summer schools organized by the Association of Constraint Programming. The event took place in the United States for the first time ever. It was hosted by UW and sponsored by the CEAS and the Artificial Intelligence Journal.

The session was geared to Ph.D. students, master's students and industrial practitioners who sought a better understanding constraint programming capabilities. There were 17 student participants, including two UW computer science undergraduates and two graduate students.

Kotthoff says the feedback from the event was very positive, and one attendee wrote, "I am highly appreciative that Professor Lars Kotthoff organized this wonderful event. I have visited many summer schools, but the quality of this one is way above average. Great lineup of speakers, great venue, and relaxed but highly stimulating atmosphere among all participants."