UAlbany Team's Language Analysis Program Key to Intelligence Efforts

Laurie Feldman

Aaron Broadwell

Researchers explore metaphors, crucial to understanding culture and language nuances

ALBANY, N.Y. (April 16, 2012) -- A University at Albany multidisciplinary research team will develop a computer-based automated program to identify and interpret language metaphors as a means of understanding culture. The program will comb through diverse text sources in four languages to understand how metaphors are deployed and apply resulting capabilities to specific, real-life scenarios. College of Computing and Information Professor Tomek Strzalkowski is leading the interdisciplinary team of researchers, which includes College of Arts and Sciences anthropologist Aaron Broadwell and psychologist Laurie Feldman.

The team will design, prototype, and test an automated language analysis system, Robust Extraction of Metaphors in Novel Data (REMND), to identify and analyze language metaphors that occur naturally in text, and apply the resulting proficiencies to real-world, analytic case studies that have significance for national security. A successful REMND system will vastly advance our capabilities of natural language understanding in cross-cultural contexts.

The 60-month REMND project is supported by a $5.5-million funding award from the Intelligence Advanced Research Projects Activity (IARPA), the research arm of the Office of the Director of National Intelligence. The REMND project is part of IARPA's Metaphor Program, based on the precept that metaphors are common and pervasive, and can reveal the underlying beliefs and worldview of members of a culture.

Tomek Strzalkowski is leading an interdisciplinary team of researchers in designing and testing the automated language analysis system, Robust Extraction of Metaphors in Novel Data (REMND). (Photo Mark Schmidt)

In Phase 1, the program will automatically discover metaphorical expressions in text, and systematically interpret their semantics in four languages -- American English, Mexican Spanish, Iranian Farsi, and Russian. In a second phase, the objective is to use these interpretations to reveal- underlying agreement and disagreement in specific instances of intercultural interaction. The project is performed in partnership with Plessas Experts Network, Inc. and Sarah M. Taylor Consulting, LLC, two small businesses with extensive expertise in language and intelligence analysis.

"The REMND project is a focused effort on the challenge of identifying and interpreting obscure and implicit meanings in language, and is a critical step in the continuing progress toward proficient and accurate analysis of multi-lingual data for the Intelligence Community," said Tomek Strzalkowski.

Head of UAlbany's Institute for Informatics, Logics and Security Studies (ILS), Strzalkowski is an expert on information analysis and natural language processing. Strzalkowski's research projects bring together computational linguistics and social science to analyze large volumes of communication and interaction in online contexts, including online games and chat rooms.

"By harnessing the theoretical and empirical research from the social sciences with the sophisticated analysis techniques from computational linguistics, we have been able to really push forward understanding about social and cultural aspects of online communication,” said Strzalkowski. In another IARPA sponsored program on Socio-Cultural Issues in Language (or SCIL), Strzalkowski’s team prototyped a highly successful approach to modeling social phenomena in discourse, including leadership and influence. “The Metaphor project will advance our research in a new and challenging direction of understanding metaphors and their use across cultures,” he added.