Reinforcement learning-based design of shape-changing metamaterials

We have implemented a new reinforcement learning method able to rationally design unique metamaterial structures, which change shape during operational conditions. We have applied this to design nanostructured silicon anodes for Li-ion batteries.