Autonomous learning of behavioral policies

Living and robotic systems share the need and the challenge of learning effective behavioral policies. Notable natural examples that will be discussed during this review talk are the tracking of surface-bound trails of odor cues, and the flight in the lowest layers of the atmosphere. These problems are relevant both biologically (animal behavior), and for robotic applications, which range from the automated location of explosives, chemical, and toxic leaks, to the monitoring of biodiversity, surveillance, disaster relief, cargo transport, and agriculture. The interdisciplinary interplay between biology, physics, robotics, and machine learning methods is key to progress and will be illustrated during the presentation.