Xi He
My name is Xi He, and I am currently a postdoctoral researcher at Peking University, working with Professor Zhenjiang Hu. I received my Ph.D. in Computer Science from the University of Birmingham, where I was advised by Professor Max A. Little. My research interests lie at the intersection of applied mathematics and computer science.
I currently study computational problems in machine learning, with broader interests in combinatorics, optimization, and algorithm design. My previous work has focused primarily on developing efficient algorithms for finding optimal (ERM) solutions to classical interpretable machine learning models. For example, I developed the first algorithms for finding optimal solutions to two-layer neural networks and hypersurface decision trees.
Research
Selected Publications
-
Deep-ICE: The first globally optimal algorithm for empirical risk minimization of two-layer maxout and ReLU networks
ICLR 2026
-
An efficient, provably optimal, practical algorithm for the 0–1 loss linear classification problem
ICLR 2026
-
Proper decision trees: An axiomatic framework for solving optimal decision tree problems with arbitrary splitting rules
Preprint; under review
-
Optimal hypersurface decision trees
Preprint; under review
Additional Publications and Preprints
-
Functional design of efficient and parallelizable combinatorial generators using convolution
Preprint; under review
-
Polymorphic dynamic programming by algebraic shortcut fusion
Formal Aspects of Computing, 2024 · Vol. 36, no. 2, Article 11
-
Foundational theory for optimal decision tree problems. I. Algorithmic and geometric foundations
Preprint