Yi-Lun Liao
My research lies at the intersection of machine learning (ML), physics, and chemistry (AI+science), focusing on modeling 3D atomistic systems. I develop algorithms by generalizing methods from natural language processing and computer vision, such as Transformers and self-supervised learning, to 3D atomistic data with physical symmetries. My works (e.g., EquiformerV1, V2 and V3, and DeNS) have consistently advanced the accuracy frontier of ML for 3D atomistic modeling, particularly as data and compute scale.
I received my Ph.D. in Electrical Engineering and Computer Science from MIT, advised by Tess Smidt. During my Ph.D., I collaborated with and interned on the Meta FAIR Chemistry team, mentored by Abhishek Das. Before that, I conducted research in ML and hardware accelerators and earned my B.S. in Electrical Engineering from National Taiwan University.