Description

The future vision for the U.S. Army includes teams of humans and intelligent agents working together to accomplish missions. DEVCOM ARL has established this new collaborative program, Strengthening Teamwork for Robust Operations in Novel Groups (STRONG), with the goal of developing the foundation for enhanced teamwork within heterogeneous human-intelligent agent teams. This collaborative venture will bring together diverse, multidisciplinary expertise to support scientific breakthroughs relevant to specific and critical scientific questions that must be addressed to enable this future vision.

STRONG focuses directly on coordination and cooperation in human-agent teams via individualized and adaptive systems. The 10-year program has a specific long-term goal to identify and implement the fundamental research necessary to develop individualized, adaptive technologies that promote effective teamwork in novel groups of humans and intelligent agents.

Prior to submitting, applicants should review Decostanza et al. (2018) Enhancing Human-Agent Teaming with Individualized, Adaptive Technologies: A Discussion of Critical Scientific Questions, which provides the initial vision and discussion of the scientific questions underlying the STRONG program, as well as Metcalf et. al. (2021) Systemic Oversimplification Limits the Potential for Human-AI Partnership, which describes re-envisioning human machine teaming paradigms to enable success in problem spaces that are un-accessible to either humans or machine intelligence alone.

Program Structure

STRONG will be executed through a series of eight annual program cycles (i.e., Cycles 1-8). The FOA will be amended annually to identify a specific problem statement, or topic, for that specific Cycle. The topic for each Cycle will be chosen to systematically converge on the specific long-term program goal.

STRONG Cycle 1: Information and Updates

STRONG Cycle 1 (FY19) was intended to set the stage with fundamental research aimed at theories of team-level processes for heterogeneous human-agent teams. Nine seedlings addressing foundational science focused on identifying and characterizing the critical states and processes for effective performance in human-agent teams.

At the end of the 2019 Summer Innovation Summit, collaborative proposals from 3 groups were requested for 3-year extensions to further explore:

1. Prediction and computational learning of human attributes, dynamics and roles for individualized agent adaptation in human-agent teams (Prime: Carnegie Mellon University)
2. Novel micro- (e.g., physiology) and meso- (e.g., social behavior) scale signatures during human-agent interaction (Prime: Northeastern University)
3. Macro-scale emergence of human-agent team processes (Prime: Northwestern University)

Two complementary, collaborative DEVCOM ARL projects were also initiated exploring coordinated physiological representations in human-agent teams and adaptive social decision making. This ongoing Cycle 1 research is focused on identifying and characterizing the critical states and processes for effective performance in human-agent teams.

Summaries from all nine seedlings and the five follow-on efforts (internal and external) can be found at the below links:

• STRONG Cycle 1 (FY19) Funded Proposal Summaries
• Adaptation of Social Decision Making under Uncertainty in Human-Agent Teams
• On a generalized framework of physiological synchrony underlying coordinated physiological representations in human-autonomy teams
• Macro Signatures of Success in Human-Autonomy Teams
• Micro and Meso Signatures of Success in Human-Autonomy Teams
• Individualized Adaptation in Human Agent Teams

STRONG Cycle 2: Information and Updates

Cycle 2 (FY20) research builds on Cycle 1 by diving deeper into understanding how micro, meso, and macro dynamics influence the emergent properties in human-agent teams, demonstrating and validating model(s) predicting human-agent team performance incorporating individual human and agent dynamics and emergent team behaviors. Nine seedlings addressing foundational science is this area were awarded.

• STRONG Cycle 2 (FY20) Funded Proposal Summaries – Summaries from all nine seedlings

At the end of the 2020 Summer Innovation Summit, 3 collaborative proposals were requested for 3-year extensions to further explore:

1. CHATS: Computational HAT model of status sensitivity to facilitate team trust and performance under suboptimal conditions (Prime: University of Delaware)
2. Physiologically Informed Adaptive Communication for Resilient Human-Agent Teaming (Prime: Colorado University, Boulder)
3. Trust-NEARCHAT: Trust Network Emergence Amongst Resource-Constrained Human-Agent Teams (Prime: University of Massachusetts, Lowell)

STRONG Cycle 3: Information and Updates

STRONG Cycle 3 focused on re-envisioning human machine teaming paradigms to optimize team performance by taking into account future team roles, capabilities, and interactions.  Projects were focused on the impact of human capabilities on rapid team reconfiguration; creating adaptable systems that offload many of the roles of humans today onto autonomy, but still are capable of leveraging uniquely human capabilities; and creating new interaction modalities that go beyond simple reward to understand human intent and co-evolve team performance. Eleven seedlings addressing foundational science in this area were awarded.

At the end of the 2021 Summer Innovation Summit, 2 collaborative proposals were requested for 3-year extensions to further explore:

1. Human-Guided ML for Futuristic Human-Machine Teaming (Prime: Drexel University)
2. The Co-evolution of Human-AI Adaptation (Prime: University of California, San Diego)

Important Program Information

STRONG addresses a critical objective within a broader Army goal to enable effective integration of Artificial Intelligence / Machine Learning (AI/ML) in the battlefield. This program has been developed in coordination with other related ARL-funded collaborative efforts (see descriptions of ARL collaborative alliances) and shares a common vision of highly collaborative academia-industry-government partnerships; however, it will be executed with a program model different than previous ARL Collaborative Research/Technology Alliances. Specific components of the program are highlighted below:

• STRONG will be executed through a series of eight annual program cycles (i.e., Cycles 1-8). The FOA will be amended annually to identify a specific problem statement, or topic, for that specific Cycle. The topic for each Cycle will be chosen to systematically converge on the specific long-term program goal.
  • Eight new topics (Cycles 1-8) are expected from FY19-FY26, with each topic focused on addressing a different scientific area within the scope of the broad research aims of STRONG. These topics will be carefully chosen based on both program achievements from the previous year and on scientific and technological advancements by the broader research community.
  • For each topic, funding will be provided to those Recipients selected for 1 year under a cooperative agreement (CA) described as the “seedling” project.
  • The Recipients of a “seedling” CA are then eligible to receive funding for a single optional extension of up to 3 years at the conclusion of the “seedling” project. The period of the performance of the option will be based on the research and available funding. It is envisioned that “seedling” Recipients will work with government researchers and/or other “seedling” Recipients to collaboratively develop a proposal for an optional extension to the initial seedling CAs. Opportunities for planning and enabling these collaborations in support of an option will be provided at the annual Summer Innovation Summits (see below), as well as via regular communication between “seedling” Recipients and government researchers.
• Recipient participation by at least one team member in each week of the Innovation Summit Series under the Recipient’s awarded Cycle will be REQUIRED.
• Proposals from junior investigators (e.g., students, research fellows, and early-career researchers with less than 5 years past reception of their PhD or less than 5 years’ experience within the primary field of their organization) are appropriate under this opportunity.

For full program opportunity details, click the Program Announcement link on this page.