About

Hi, I’m Winnie, an undergraduate studying Chemical Engineering and Computer Science at National Tsing Hua University (NTHU), Taiwan.

My interests revolve around representation learning that respects the symmetries of nature. Recently, I’ve been particularly drawn to neural quantum states and variational modeling of quantum many-body systems beyond the practical reach of conventional ab initio methods, such as strongly correlated systems.

At Caltech’s Anima AI+Science Lab, I’m part of the AI for chemistry group, advised by Prof. Anima Anandkumar. My latest work focuses on SU(2)-equivariant graph neural networks for molecular magnetism. I also helped with GPU acceleration of the OrbitAll foundation model. Currently, at NTHU’s Lab for Material and Molecular Design, I’m developing equivariant GNNs for vibrational circular dichroism (VCD) spectroscopy with Prof. Kun-Han Lin. My earlier work there includes building an automated force-field parameterization toolkit and studying high-entropy alloys for the hydrogen evolution reaction using density functional theory.

I’m also an LLM Research Intern (Part-time) at the National Center for High-Performance Computing (NCHC). Previously, as captain of the NTHU Student Cluster Competition Team, I led NTHU teams at national, Asia-Pacific, and international competitions. More details are in my CV.

Outside research, I’m a cat lover, and I enjoy playing chess and piano. Feel free to say hi!