Exact schedule: 11:15-11:55
Imagine how crumpling a piece of paper creates persistent folds. Deforming a soft material often results in lasting changes. In conventional materials, these changes are detrimental; folds and cracks accumulate until the material fails. But living systems can improve when challenged, as does muscle with exercise or the brain with study. I will present an artificial metamaterial that becomes better at executing a robotic motion through repeated practice. Our material is inspired by the plasticity rules that determine when neurons form connections in the brain. We translate these rules into mechanical networks of variable-length linkages acting as trainable synapses, enabling the material to learn. I will also discuss my research plans to understand how computation emerges from the properties of its material substrate, and to make novel active and learning materials at the microscale.