Amin Heyrani Nobari
Amin Heyrani Nobari is a PhD candidate whose research focuses on the development of novel machine-learning (ML) models and methods for automating CAD generation. His current work aims to develop generative adversarial networks (GANs) for design generation and conditional GANs for inverse design problems. Promising results of Amin’s research to date include a new GAN- based inverse design synthesis method named PcDGAN, which outperformed state-of-the-art machine-learning methods by 69%, and a novel algorithm for generating creative designs called CreativeGAN. As a MathWorks Fellow, Amin will continue to pursue new approaches in inverse design of linkage mechanisms and methods that enable algorithms to propose 2-D planar linkage mechanism designs for any given target. To enable his graph neural network ML models, Amin has developed a 10x faster algorithm to simulate mechanisms, and he publicly released a dataset of 100 million mechanisms—a resource that could be extremely valuable for researchers working on large-scale, deep learning-based linkage mechanism design.