PhD Student


Department of Integrative Biology
University of California at Berkeley
3060 Valley Life Sciences Building
Berkeley, California 94720-3140

Lab phone: 510-643-5183
Fax: 510-643-6264

Email: nathaniel.hunt [at]


Learning allows animals to extend their behavioral repertoire beyond the innate behaviors hard-coded in their neuromechanical systems by evolution.  While many mobility problems can be solved by stable and robust design, sensorimotor learning can extend these abilities allowing animals to successfully navigate in especially complex, unforeseen and changing environments.  Similarly, learning control in robotics will be a critical complement to stable, robust design in bringing robots out of the laboratory and into the complex, dynamic and often unpredictable real world, allowing them to grow an ever more sophisticated constellation of un-programmed behaviors.  I am interested in how learning explores and exploits the informational and mechanical state space to generate task-specific control heuristics that work synergistically with the reduced dimensional template dynamics of the animal.

Quantifying multiple sequential trials of small, legged, terrestrial animals executing goal oriented maneuvers in rich environments will allow us to extract learning principles regarding the use of proprioceptive and exteroceptive sensory feedback information to produce highly skilled movements.  Identifying the principles of how these hierarchical modular behaviors form, organize and integrate will inform the design of improved bio-inspired robotic designs and their learning algorithms.  I believe this problem is best approached in a highly multi-disciplinary manner using techniques from biology and biomechanics to machine learning, and nonlinear dynamical systems.