Research Interests

Interdisciplinary studies of artificial intelligence and science (AI for Science), including bioinformatics, medical image analysis, cheminformatics, and the design of new materials.

Research Topics

  • Molecular Representation Learning: Developing deep learning, self-supervised learning, and large language models to decipher the properties, functions, and interactions of compound molecules and biological macromolecules. Our methods are further extended to predict chemical reaction outcomes and facilitate retrosynthesis, aiming to speed up drug discovery and materials design.
  • Biomedical image understanding: Designing semi-supervised and multi-modal learning approaches for pattern recognition and segmentation of biomedical images, with applications to biomarker identification and disease outcome prediction.

News

2024

DateNews
Aug 19-22Lab members attended ICPOC 2024. Gufeng and Letian presented their posters on LLM for retrosynthesis and molecular chiral representation, respectively.
Aug 6-7Prof. Yang attended the 2nd RIKEN AIP-SJTU CS Joint Workshop on Machine Learning and Brain-like Intelligence (Agenda) at RIKEN AIP, delivering a talk on Uncertainty-based Semi-supervised Learning in Medical Image Segmentation.
July 19-21Lab members attended ISBRA 2024. Yifei and Yujia delivered oral presentations on their works on RNA-protein binding and RNA localization prediction, respectively.
July 5Lab members attended ACM TURC 2024, where Prof. Yang gave an invited talk "Deep Learning Revolutionizes Chemical Informatics: Model Evolution and Applications".
June 16Lab members attended the 2024 Bioinformatics and Intelligent Information Processing Conference (BIIP2024), where Prof. Yang and Runhan presented on "LLM in Chemistry" and "Chemical Reaction Prediction via Multi-view Learning," respectively.
May 30Our database CLC-DB containing 1800+ chiral ligands/catalysts of different chiral types is online!
May 30Yichong and Wenjing passed their PhD proposal defense. Congratulations!
May 24Prof. Yang gave a talk "Medical Image Analysis Driven by Large Models: From Single-Modality to Multi-Modal Learning" at the 13th Shanghai Jiao Tong University Vascular Disease Forum.
May 16Our collaborative work "Measuring Bargaining Abilities of LLMs: A Benchmark and A Buyer-Enhancement Method" is accepted accepted to the ACL Findings. Congratulations!
May 7Our proposed workshop, "Large Foundation Models for Multi-Modal Learning in Bioinformatics and Biomedicine Informatics," has been officially accepted for inclusion in BIBM'2024. We warmly invite researchers, academics, and industry professionals to contribution original works to our workshop!
Apr 17Yuyang's work ClusterDrop for addressing over-smoothing issue in GNNs is accepted by IJCAI 2024. Congratulations!
Apr 15Our GaCaMML paper is accepted by Artificial Intelligence in Medicine. Congratulations, Yuzhang and Guoshuai!
Mar 29Zebei, Guoshuai, Kuo, and Jingyao completed their Computer Science MS degree. Congratulations!
Feb 25Runhan's work RMVP for large-scale prepretraining of chemical reactions is published in Journal of Cheminformatics. Congratulations!
Feb 23Our CGMega paper is accepted by Nature Communications. Congratulations, Zebei!
Feb 22Our collaborative work CircSite on predicting circRNA-protein binding is published in Computers in Biology and Medicine. Congratulations!
Feb 10Guoshuai's work for designing new Al-Zn-Mg-Cu-Zr-Hf alloy is published in Materials Today Communications. Congratulations!