My main research area lies in machine learning, with special interests in models and algorithms inspired by theoretical insights. Recently, my works focus on representation learning on structured data (e.g., graphs and sequences) and provide comprehensive analysis on base architectures (Transformers, GNNs) covering their expressiveness, efficiency and effectiveness, as well as apply them to AI for Science, Graph Learning and Natural Language Processing. I have published several papers and been reviewers at the top machine learning and artificial intelligence conferences such as NeurIPS, ICML and ICLR.