ESPCI web site
Programming materials with classical conditioning algorithms
There exists a growing need for adaptive functional materials. Could materials showing a response to a particular stimulus become responsive to another stimulus to which they are originally indifferent ? Such behavior would mimic classical Pavlovian conditioning in behavioral psychology, one of the elementary forms of learning. This would conceptually differ from conventional stimuli-responsive and shape memory materials, where no new stimuli become activated to allow the same response. Here we demonstrate soft matter systems programmed to show classical conditioning, allowing to respond to an initially neutral stimulus (light) after a conditioning process, where the neutral stimulus is associated with an unconditioned stimulus (heat). Inspired by the biological processes, we also show "forgetting" and "recovery" by out-of-equilibrium processes, incorporating tools from systems chemistry. We suggest generalizations for other stimuli and materials emulating algorithmically the elementary mechanistic aspects of learning.