Our group studies systems that are inherently disordered (e.g., glasses, packings, driven suspensions, crumpled sheets). Using computational and theoretical techniques we strive to understand how simple microscopic rules give rise to complex emergent behaviors. A central theme of our research is the idea of self-organization. A system attains a function or structure from the inter-particle interactions, rather than a driving hand that controls the many degrees of freedom. Recently, we have been using these ideas to “train” materials with non-trivial elastic responses building on analogies with learning theory. This enables materials that perform computations or classification much like neural networks.