Yu Zhu

Yu Zhu

AI for Science · Computational Chemistry · Computer Vision

About

Yu Zhu works at the intersection of AI for Science, computational chemistry, and computer vision. His research spans machine learning force fields, physics-driven reconstruction of molecular orbitals from scanning tunneling microscopy images, open-set fine-grained recognition, and materials modeling.

Education

M.Sc., Department of Chemistry

2023–2026

Fudan University

B.Sc., College of Chemistry, Chemical Engineering and Materials Science

2019–2023

Soochow University

Research Interests

AI for Science Machine learning force fields Scanning tunneling microscopy Open-set recognition
Research

My research connects physical modeling with modern machine learning. I develop physics-driven methods for reconstructing molecular-orbital information from scanning tunneling microscopy images, and I work on machine learning force fields and scalable materials-modeling workflows.

I am also interested in open-set fine-grained recognition, where models must recognize known categories while detecting previously unseen ones.

Featured Publications
Recent Publications
Research Projects

MPBR for Open-Set Fine-Grained Recognition

Multimodal progressive bidirectional reasoning for suppressing semantic drift in open-set fine-grained recognition.

MatterSim: Materials Modeling Workflows

Benchmarks and scalable computational workflows for a large-scale pretrained deep learning atomistic model.

STM-Net: Reconstructing Molecular Orbitals from STM Images

Physics-driven STM image restoration with PSNR above 37 dB using 10% of the training data and above 43 dB on the full dataset.

SERS Imaging for Fingerprint Identification

Surface-enhanced Raman scattering imaging on flexible substrates for chemical mapping of latent and live fingerprints.