Science of Understanding: Modeling Behavior and Learning in Humans and Machines

 

We envision a future where humans and machines are training and collaborating in natural ways to accomplish mission objectives. To develop the rich mutual understanding needed between collaborative teammates, we are pursuing the science of understanding, developing formal computational and mathematical models of human learning and decision making and integrating these models with machine intelligence to close the representational and communication gaps between human and machine teammates. We seek to advance our scientific knowledge of how perceptual, cognitive and motor processes come together to shape human behavior and understanding of both the world and their teammates, both human and synthetic. And we seek to advance our scientific knowledge about how human cognition, behavior, and understanding are shaped by experiences and interactions with artificial intelligence systems and synthetic teammates. We are further interested in how interactions with humans can advance the intelligence of synthetic teammates and machine learning systems.

 

We seek research in several related areas: (1) advancing formal computational and mathematical models of cognition that deepen our understanding of human perception, cognition and decision making, (2) methods for leveraging cognitive models for measurement (cognitive psychometrics), assessment and real-time inference, (3) novel approaches and methods for interactive learning between humans and machines, (4) advances in the robustness of modeling and simulation technologies supporting human-machine team training, and (5) methods for test, evaluation, validation and verification of cognitive models and interactive training paradigms. Our team leverages approaches that combine human subjects experimentation, formal theory development, cognitive modeling, machine learning, visual analytics and cognitive systems engineering. To meet our goals of advancing both scientific knowledge and mission-relevant capabilities, we leverage laboratory studies that isolate phenomena of interest and complex synthetic task environments inspired by real-world USAF operational requirements.

 

Research areas of interest include, but are not limited to, mathematical psychology, cognitive modeling, human factors, visual analytics, active and interactive machine learning, artificial intelligence, and human-computer interaction.

 

 

References

Blaha, L. M. (2018, May). Interactive OODA processes for operational joint human-machine intelligence. NATO IST-160 Specialist’s Meeting: Big Data and Military Decision Making. Bordeaux, France.

Blaha, L. M., Lebiere, C., Fallon, C. K., & Jefferson, B. (2020). Cognitive mechanisms for calibrating trust and reliance on automation. Proceedings of the 18th International Conference on Cognitive Modeling.

Fallon, C., Blaha, L. M., Cook, K. & Billow, T. (2019, July) Common ground and autonomy: Two critical dimensions of a machine teammate. 10th International Conference on Applied Human Factors and Ergonomics. Washington, D.C.

 

 

key words

Cognitive modeling; Mathematical modeling; Computational models; Human performance; Human learning; Decision making; Cognitive science; Machine learning; Artificial intelligence; Human-autonomy teaming; Understanding; Common ground

Citizenship:  Open to U.S. citizens

Level:  Open to Postdoctoral and Senior applicants