📍 SLDS 2026 — I will be attending SLDS 2026 in New York (Nov. 1 – Nov. 3), where I will present my work Transformers as Provable Approximators of Sparse Principal Component Analysis in the poster session. Feel free to say hi if you’ll be there!
About me
I am a senior at School of the Gifted Young, University of Science and Technology of China (USTC), majoring in Mathematics with a concentration in Probability Theory and Mathematical Statistics.
I am applying for Ph.D. positions starting in Fall 2027. My research interests lie broadly in statistical machine learning and optimization, with an emphasis on the theoretical foundations of modern learning architectures and on how structure can guide model analysis and design. My current work focuses on understanding Transformers as algorithmic learners for unsupervised spectral estimation, and on diffusion-based generative modeling for structured data.
Throughout my research, I have been fortunate to work closely with several wonderful mentors and collaborators. At USTC, I am advised by Prof. Canhong Wen on understanding how Transformers can perform fully unsupervised high-dimensional spectral estimation, and I collaborate with Dr. Yan Chen on this line of work. I am also advised by Prof. Ji Zhu and Prof. Gongjun Xu at the University of Michigan, and work closely with Prof. Shihao Wu at UC Davis on generative modeling for hypergraphs with latent hyperbolic representations. Their guidance and encouragement have shaped both my research and the way I approach scientific questions, and I am deeply grateful to have learned from and worked with them.
My name can be pronounced as “cheen-yen lyoh” and my name in Chinese is 刘沁妍.
Contact & Links
- Email: qinyanliu2006@gmail.com
- CV: PDF
- Scholar: Google Scholar
- GitHub: github.com/qinyanliu
Education
2023.9 - 2027.6 (Expected) School of the Gifted Young, USTC
2021.9 - 2023.6 High School Affiliated to Nanjing Normal University
2018.9 - 2021.6 Nanjing Foreign Language School