Hi there! My name is Xi Wang. Currently, I am a research assistant at the Courant Institute, New York University, working under the supervision of Shengjie Wang. My research interests lie in developing innovative machine learning models to tackle scientific challenges, particularly in the fields of biology and chemistry. I’m also interested some classic topics in machine learning, such as explainability and hallucination.

I am actively seeking Ph.D. opportunities in the field of AI for Science. Please feel free to reach out if you have any suggestions or opportunities to share!

Research Interest

AI for Science

My work focuses on applying AI to biology and organic chemistry. Specifically, I am interested in addressing the hallucination problem in deep learning models. In AI for Science research, hallucination can be particularly detrimental, as it often violates the moral, physical, or chemical principles that govern these domains. My goal is to improve the reliability of AI systems in scientific applications.

Computational Biology

Computational biology is a complex discipline that focuses on using computational methods to solve problems in biological sciences. My work specifically centers on biological macromolecules, with a focus on the structure and function prediction of proteins, RNA, and antibodies. I utilize pre-trained models to enhance the predictive accuracy and generalization performance in downstream tasks.

Molecules Design

Chirality is a fundamental property of nature, as it influences both the proteins that make up living organisms and the drugs that cure diseases. My research focuses on a special class of chiral molecules known as geometrically chiral molecules. Unlike centrally chiral molecules, these molecules create a chiral environment by forming asymmetric three-dimensional “pockets”. Such chiral molecules have significant applications in pharmaceuticals and catalysis. I integrate quantum chemical calculations, pre-trained models, and diffusion models to develop new models for addressing related challenges.