TechConnect Innovation Program

Socially Intelligent Virtual Agents for use Training, Games, and Customer Relations

SIFT - Smart Information Flow Technologies, MN, United States

TECHNOLOGY SUMMARY

A computational model of interactional sociolinguistics gives social intelligence to otherwise empty, emotionless virtual agents. Virtual characters that act appropriately (e.g. empathetic, offended) increase the interest and engagement of the human, whether for training scenarios in military simulations, video games with compelling characters, or as virtual help desk attendants.

Primary Application Area: IT, Cyber-Security, Software

Technology Development Status: Prototype

Technology Readiness Level: TRL 6

FIGURES OF MERIT

Value Proposition: Our technologies will allow you to enhance training and customer relations through intelligent, helpful, and likeable employees who can connect, who happen to be digital.

SHOWCASE SUMMARY

Organization Type: Academic/Gov Lab

Showcase Booth #: 713

Website: www.sift.net

GOVT/EXTERNAL FUNDING SOURCES

Government Funding/Support to Date: SIFT has been at the forefront of optimizing human-automation with linguistics since 2001. This work has continually evolved over the last 15 years, resulting in many research efforts and applications in different domains. In 2001 we designed automation-generated reminders for a medication reminder system, part of a larger $5M effort funded jointly by Honeywell and NIST. The computational model evolved to include social dynamics in a DARPA SBIR, which was applied to virtual agent interactions in a cultural training tool for soldiers. In a US Air Force sponsored SBIR, we validated this model and quantified its impact on human performance. Now called the SIFT Etiquette EngineTM, we applied the technology to unobtrusively identify team dynamics such as power structures in a large military exercise at Ft. Huachuca through the use of chat data in an ONR BAA. Most recently, under 2 NASA funded projects, we expanded the Etiquette Engine work to perform extensive validation through three NASA test facilities, and are creating virtual agents to support NASA’s vision of future manned long duration exploration missions.

Primary Sources of Funding: Federal Grant, Other

Looking for: Both Funding and Development Partners