Aptima Receives DARPA Award
DARPA Awards Contract to Aptima & ASU for Realistic, Collaborative Agents for Human-AI Teaming
ADAPT borrows from human cognition to offset challenges of modern battlespace
Aptima has announced a multi-year contract award from the Defense Advanced Research Projects Agency (DARPA) for ADAPT, Adaptive Distributed Allocation of Probabilistic Tasks. ADAPT is an innovative program to develop a new generation of AI and (software) agents designed to work alongside, learn from, and interact with human teams, helping automate, plan, and execute missions for the dynamic speed and uncertainty of modern military operations missions.
As adversaries become more adept and capable, the future success of military teams operating in fast-changing battlespaces will depend on accelerating the observation-to-action loop. AI will be critical in assisting commanders with tactical and strategic decision-making, however, current AI and agents are limited in their reasoning and integration with human teams, reducing effectiveness in planning and action cycles.
Under the ADAPT contract, Aptima and partner Arizona State University (ASU) will design, build, and validate collaborative, adaptive, and realistic AI agents. These innovations will advance the state of human-machine teams by not only improving communications and cooperation between human and artificial agents, but by using AI to help commanders process streams of real-time data to structure teams, and to create and adapt action plans as mission conditions change.
Borrowing from Humans
ADAPT’s agents and AI systems are inspired by qualities of human cognition, using a type of reasoning based on ‘active inference.’ This will enable synthetic agents to perceive, learn, and adapt based on their observations of the environment and their interactions with human team members. ADAPT will address another critical limitation in current systems by enabling the bi-directional communications and inputs that are needed between humans and agents for collaboration, using a variety of advanced interfaces, visualizations, and models.
Adam Fouse, Ph.D. Director, Performance Augmentation Division and ADAPT Program Manager
“ADAPT will take a significant step forward in human-AI collaboration so Warfighters and intelligent technology can reason and work together to make better, faster decisions than either could do on their own,” said Dr. Adam Fouse, Aptima’s ADAPT Program Manager. “By learning from its human counterparts, and taking into account their goals, preferences, and constraints, these more informed agents can guide AI in forecasting, creating, and adapting action plans as missions evolve.”
For example, in an urban search and rescue scenario that involves searching a smoke-filled complex, these advanced AI models and agents would reason over millions of possible scenarios to help identify what resources were needed where, helping commanders to enact the best plan to deploy various first responders, robotic, and autonomous assets, while minimizing casualties and risks from a possible building collapse.
“Humans excel at learning from one another but can only process so much incoming information. AI on the other hand has incredible computational abilities but requires the ability to learn from and communicate with humans in order to be used effectively in dynamic team situations,” Fouse added. “These combined attributes will elevate a commander’s expertise and decision making in fast-changing, information-intensive environments so they can respond, and adapt quickly, while also considering future possibilities.”
The Mission Challenge
On the battlefield, where the notion of a static battle plan is obsolete, US military forces must operate with a strategy that takes into account this influx of information. Planning, however, becomes too complex for humans to identify the optimal strategy for dynamically reallocating assets on the fly while also considering uncertain future requirements. “As an advanced decision support system, ADAPT will learn from the environment and other team members, using its AI models to formulate best actions plans to assist its human counterparts,” Fouse added.
ADAPT is closely tied to two ongoing DARPA AI programs: Agile Teams (A-Teams), for which Aptima is the prime contractor, and Artificial Social Intelligence for Successful Teams (ASIST), where Aptima is priming the evaluation effort. ADAPT is extending Aptima’s work in A-Teams in solving the challenge of distributed decision making in changing environments, and how human-machine teams dynamically allocate tasks, abilities, and workloads across a blended operation of human, autonomous, and networked systems. For ASIST, ADAPT’s resulting AI and agents will be tested in human-in-the-loop experiments in the ASIST testbed, which uses a Minecraft-based urban search-and-rescue synthetic task environment (USAR-STE) recently developed by Aptima and ASU.
About Aptima, Inc.
For 25 years, Aptima’s mission has been to improve and optimize performance in mission-critical, technology-intensive settings. We apply deep expertise in how humans think, learn, and perform to today’s challenges. Whether for fighter pilots functioning in the cockpit, medical staff in the ICU, or teams collaborating across distributed networks, our solutions help measure, assess, inform, and augment human performance in defense, intel, aviation, law enforcement, and healthcare. For more information, visit www.Aptima.com.
Media Contact:
Joel Greenberg | DCPR
joel@dcpr.com | 202-363-1065 | 202-669-3639 cell
DISTRIBUTION STATEMENT A. Approved for public release: distribution unlimited.