Research Topics

• Biological Language Modeling

We regard biological sequences (DNA/RNA sequences and amino acid sequences) as a special kind of language, and explore biomolecular structure and functions based on sequence analysis using large language models and deep learning methods.

• Biomedical Image Understanding

We design machine learning algorithms for the annotation, clustering, and segmentation of biomedical images, including the recognition of complex molecular localization patterns in biological microscopy images and the identification of biomarkers in medical imaging.

• Representation Learning of Molecules and Materials

We extract features from compound molecules and chemical reactions, propose novel descriptors for elements and materials, and utilize these for predicting chemical reaction outcomes, retrosynthesis, and the design of new materials with enhanced properties.

• Machine Learning/Deep Learning Model Research

We focus on the uncertainties in machine learning models, including the modeling, measurement, and mitigation of epistemic and aleatoric uncertainties.