Published several papers, such as 'Unsupervised Transfer Learning via Adversarial Contrastive Training' and 'Distribution Matching for Self-Supervised Transfer Learning.' Maintains an open-source machine learning implementation library called mlimpl, which provides concise and readable implementations of popular algorithms, suitable for teaching, prototyping, and reproducible research.
Research Experience
Engaged in multiple research projects, including unsupervised transfer learning via adversarial contrastive training and distribution matching for self-supervised transfer learning.
Education
Currently pursuing a PhD degree in the Department of Applied Mathematics at Hong Kong Polytechnic University, under the supervision of Prof. Defeng Sun and Prof. Houduo Qi. Previously, obtained a Bachelor’s degree from Northwest University and a Master’s degree from Wuhan University, with Master’s supervisor Prof. Yuling Jiao.
Background
Current research interests lie in statistical learning and deep learning theory, particularly in deep representation learning, self-supervised learning, and the theoretical foundations for large language models (LLMs). Focuses on establishing theoretical understandings and developing novel approaches in these areas.