Computer Science Events
CS Homecoming Colloquium: Eric Davis, DARPA
Access & Features
- Open to the public
Title: "Third Wave Artificial Intelligence, Human Machine Teams, and Crisis Reasoning"
In order for artificial intelligence to function as an artificial teammate rather than decision support, it must support and understand interdependent workflows with human co-performers. Interdependence requires an understanding of the current state of the shared task environment as well as what that state means to co-performers, even when that meaning is defined by subjective opinions. Human co-performers embedded shared concepts into lower dimensional manifolds that support Euclidean distance measures allowing the communication of intuitive notions of "closeness" or distance between concepts related to joint activity. Effective machine teammates must replicate this process.
We will explore the critical role of human-machine teaming within crisis reasoning scenarios and how it impacts performance in high-stakes environments, delving into the concept of self-synchronization—a fundamental aspect of team intelligence where mutual understanding and synchronization between human teammates can lead to superior outcomes. However, failures in this area, such as those observed in air traffic control and anesthesiology, highlight the challenges posed by working memory limitations and attention tunneling for humans in crisis scenarios. We will examine the theoretical underpinnings of self-synchronization, its practical implications, and the emerging models from multidisciplinary science that seek to bridge gaps in team knowledge. Special attention will be given to how geometric reasoning and intuitive knowledge representation contribute to effective collaboration and how these principles are being integrated into AI systems to enhance team performance. The talk will also address the future trajectory of human-machine teaming with Third Wave AI, focusing on the development of models that facilitate better alignment and communication between human and machine agents.
Bio:
Dr. Eric Davis joined the Defense Advanced Research Project Agency's (DARPA) Strategic Technology Office (STO) in 2024. His research interests include artificial intelligence, cognitive science, and human-machine teaming. Prior to DARPA, Davis was the principal scientist of artificial intelligence, machine learning, and human-machine teaming at Galois, and a professor and director of the Trustworthy Data Engineering Laboratory at Iowa State University and the University of Miami. He was named a Frontiers of Engineering Education faculty member by the National Academy of Engineering, earned his bachelor’s in Computer Science from William & Mary and his PhD from the University of Illinois at Urbana-Champaign as an IBM doctoral fellow.
Sponsored by: Computer Science
Contact
Huajie Shao