Leif Johnson

Mountain View College, The University of Texas at Austin

Papers

Total Citations

H-Index

About

Leif Johnson is a researcher whose work spans computational modeling of human behavior, autonomous systems, and artificial intelligence. His research bridges cognitive science and machine learning, with a particular focus on how intelligent agents — both human and artificial — navigate uncertainty and adapt their behavior in complex environments. Johnson's most notable recent contribution, "Resolving Uncertainty on the Fly: Modeling Adaptive Driving Behavior as Active Inference" (2024, 18 citations), applies the active inference framework to model how human drivers dynamically manage uncertainty while driving. This work addresses a critical gap in autonomous vehicle development, where realistic simulated human driver models are essential for safely evaluating self-driving systems. By grounding adaptive driving behavior in principled probabilistic theory, Johnson's approach offers a more cognitively plausible alternative to existing traffic psychology models. His earlier work, "Multiagent Learning through Neuroevolution" (2012, 13 citations), demonstrates his longstanding interest in how multiple learning agents can develop intelligent behavior through evolutionary computational methods — a foundational contribution to the multiagent systems literature. Across his career, Johnson's research consistently tackles how agents learn, adapt, and make decisions under uncertainty, making his work highly relevant to researchers in autonomous systems, cognitive modeling, and artificial intelligence.