Xuechen Li

University of Toronto

Biography

I am a brain resident at Google.

I am interested in computational aspects of statistics and optimization, with a focus on methods and techniques that interact with machine learning. I aim to improve our understanding of modern machine learning with tools from applied maths and use the derived insights to develop more efficient and reliable algorithms.

I am most experienced with Monte Carlo methods and numerical analysis.

I will be maintaining a blog here.

Interests

Statistics
Optimization
Machine Learning

Education

BSc in CS, Math & Stats, 2019

University of Toronto

Research

Xuechen Li, Ting-Kam Leonard Wong, Ricky T. Q. Chen, David Duvenaud. Scalable Gradients for Stochastic Differential Equations. In submission, 2019.

Xuechen Li, Denny Wu, Lester Mackey, Murat A. Erdogdu. Stochastic Runge-Kutta Accelerates Langevin Monte Carlo and Beyond. Advances in Neural Information Processing Systems (NeurIPS), 2019.

PDF Spotlight

George Barmpalias, Neng Huang, Andrew Lewis-Pye, Angsheng Li, Xuechen Li, Yicheng Pan, Tim Roughgarden. The idemetric property: when most distances are (almost) the same. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2019.

PDF

Ricky T. Q. Chen, Xuechen Li, Roger Grosse, David Duvenaud. Isolating Sources of Disentanglement in Variational Autoencoders. Advances in Neural Information Processing Systems (NeurIPS), 2018.

PDF Oral

Chris Cremer, Xuechen Li, David Duvenaud. Inference Suboptimality in Variational Autoencoders. International Conference on Machine Learning (ICML), 2018.

Preprint

Contact

lxuechen@cs.toronto.edu