Publications: Preprints like 'Coevolutionary Continuous Discrete Diffusion' and 'On Powerful Ways to Generate: Autoregression, Diffusion, and Beyond'; 'Next Semantic Scale Prediction via Hierarchical Diffusion Language Models' accepted to NeurIPS 2025; 'Learning Diffusion Models with Flexible Representation Guidance' accepted to NeurIPS 2025 and presented as Oral at ICML 2025 FM4LS Workshop; Graduated from Tsinghua University with honors, awarded Outstanding Graduate (Top 1.5%) and Outstanding Undergraduate Thesis by Beijing Ministry of Education and Tsinghua University.
Research Experience
Summer 2025 internship at Microsoft Research, working with Dinghuai Zhang; Summer 2023 internship at UCSD, advised by Yusu Wang and Rose Yu; worked with Gao Huang on computer vision at Tsinghua University.
Education
PhD: Massachusetts Institute of Technology, EECS, co-advised by Tommi Jaakkola and Stephen Bates; Bachelor's: Department of Automation, Tsinghua University, minor in Statistics, worked with Muhan Zhang at the Institute for Artificial Intelligence, Peking University.
Background
Research interests: theoretical and applied machine learning, generative models, AI for Science. Areas of focus include statistics, expressivity, optimization, diffusion models, LLM reasoning, multi-modal foundation models, geometric deep learning, AI4biology, and AI4healthcare.