Prof. Dragomir Radev helps coach U.S. linguistics team to victory   Bookmark and Share

Prof. Dragomir Radev participated as the U.S. team coach in the Fifth International Linguistics Olympiad, held recently in St. Petersburg, Russia. This is the first year that teams from the U.S. competed, and they came home with several victories.

The International Linguistics Olympiad has its origins in the Olympiad of Linguistics and Mathematics, founded in 1965 in Moscow. High-school students compete by solving linguistics and logic problems based on natural language. Eight finalists from the 200 students who participated in the North American Computational Linguistics Olympiad earned the honor to compete in the international competition; this is the first time U.S teams participated.

USA’s Team 2 won the team contest (tied with a Russian team); members of the team were Rebecca Jacobs of Los Angeles, CA, Michael Gottlieb of Dobbs Ferry, NY, Josh Falk of Pittsburgh, PA, and Anna Tchetchetkine of San Jose, CA.

Individual members of USA’s Team 1 Adam Hesterberg of Seattle, WA, won the individual contest, and Jeffrey Lim of Arlington, MA, won a Best Solution award. The U.S. teams were sponsored by Google, the National Science Foundation, and the North American Chapter of the Association for Computational Linguistics (NAACL).

Prof. Radev is a natural as a coach, having won several awards himself as a high school student in Bulgaria back in the 80’s, when there was no international round. In addition to being coach for the international contest, Prof. Radev was program chair for the U.S. contest.

“There are a number of connections between this contest and my own research,” said Radev. “The closest one is to my research in machine translation. In machine translation, one of the main problems is to learn rules from aligned sentences in two languages and then using these rules to translate new sentences from one of the languages to another. Additionally, solving the linguistics problems involves a number of Artificial Intelligence techniques such as constraint satisfaction and search.”

Additional Information