DARPA picks SRI International to develop natural language computer processing to bridge cultural differences

Jan. 11, 2022
SRI to develop natural-language processing that adapts to and recommends how to operate within different emotional, social, and cultural norms.

ARLINGTON, Va. – U.S. military researchers are asking SRI International in Palo Alto, Calif., to develop natural language computer processing that could help military forces understand and adapt to different emotional, social, and cultural norms that exist throughout the world.

Officials of the U.S. Defense Advanced Research Projects Agency (DARPA) in Arlington, Va., announced a $10 million three-year contract to SRI International last month for the Computational Cultural Understanding (CCU) program.

For the CCU project, SRI computer scientists will develop natural language processing that recognizes, adapts to, and recommends how to operate within the emotional, social, and cultural norms that differ across societies, languages, and group affinities.

U.S. military counter-insurgency planners are in constant contact with diverse cultures, which requires communicative understanding, not simply of local languages, but also of social customs and cultural backgrounds, DARPA researchers say.

Related: Blending different sensors, machine learning, and artificial intelligence (AI) is aim of DARPA research

This can involve targeting feelings as well as ideas. Cross-cultural miscommunication not only derails negotiations, but also can be a contributing factor leading to war, experts point out. The likelihood of communications failure increases dramatically where significant social, cultural, or ideological differences exist.

The CCU program asks SRI International to create cross-cultural language understanding technologies to improve military situational awareness. These new natural language processing technologies should require minimal-to-no training data in a local culture.

Instead of relying primarily on annotated training data, SRI International will capitalize on qualitative and quantitative findings from fields such as psychology, sociology, and related disciplines in minimally supervised machine-learning techniques to infer the meaning of unlabeled discourse behaviors in context.

Related: DARPA launches DEFT program to pull actionable intelligence out of ambiguously worded text

Today's automated data processing systems are incapable of analyzing cross-cultural communication accurately, or providing useful assistance to negotiations beyond basic machine translation.

SRI International experts will try to develop computer technology that can achieve relative parity with human interpreters, who provide cultural insight and leverage this knowledge in the translation process, and help negotiators and analysts with language analysis and cross-cultural dialogue in the field.

CCU technology will help automatically to discover sociocultural norms to determine the social, cultural, and contextual factors to help with communications and rapport building. The project also will help military planners recognize emotions across different languages and cultures.

Related: DARPA releases solicitation for artificial intelligence job to advance machine learning

SRI International will develop new component technologies, and incorporate these technologies into a prototype for cross-cultural dialogue assistance, which will create the foundation for a fieldable capability.

The CCU project will focus on computer-aided sociocultural analysis; sociocultural norm discovery; cross-cultural emotion recognition; communicative change detection; and data creation for development and evaluation.

For more information contact SRI International online at www.sri.com, or DARPA at www.darpa.mil.

Voice your opinion!

To join the conversation, and become an exclusive member of Military Aerospace, create an account today!