Metamaterials are a novel category of synthetic materials that exhibit unique properties that are not found in nature materials. Their exceptional properties are achieved by meticulously designing their microstructures. However, traditional design approaches are computationally expensive, limiting the range of materials that can be designed. In this project, we will use advanced generative models (such as generative adversarial network, denoising diffusion model) to design metamaterials with specific desired properties. The ideal candidate should possess of some basic knowledge about mechanics and artificial neural networks.