I am a third year PhD Student at The Operations Research Center at MIT, where I am fortunate to be advised by Rahul Mazumder.

Research Interests

I am interested broadly in the use of mathematical optimization methods (first/second-order methods, mixed-integer programming, robust optimization) in deep learning and high-dimensional statistics.

Papers

A GPU-accelerated Nonlinear Branch-and-Bound Framework for Sparse Linear Models arXiv
Xiang Meng, Ryan Lucas, Rahul Mazumder

Reasoning Models Can Be Accurately Pruned via Chain-of-Thought Reconstruction arXiv
Ryan Lucas, Kayhan Behdin, Zhipeng Wang, Shao Tang, Qingquan Song, Rahul Mazumder Workshop paper at NeuRIPS 2025, First Workshop on Efficient Reasoning
Conference paper at International Conference on Learning Representations (ICLR, 2026)

Pixie: Fast and Generalizable Supervised Learning of 3D Physics arXiv, Website
Long Le, Ryan Lucas, Chen Wang, Chuhao Chen, Dinesh Jayaraman, Eric Eaton, Lingjie Liu

Preserving Deep Representations in One-Shot Pruning: A Hessian-Free Second-Order Optimization Framework arXiv
Ryan Lucas, Rahul Mazumder
Conference paper at International Conference on Learning Representations (ICLR, 2025)

Certified Robust Neural Networks: Generalization and Corruption Resistance arXiv
Amine Bennouna, Ryan Lucas, Bart Van Parys
International Conference on Machine Learning (ICML, 2023)
– Runner-up of INFORMS 2023 Data Mining Best Paper Award (General Track)
– Winner of INFORMS 2023 Workshop on Data Science Best Student Paper

Holistic Robust Data-Driven Decisions arXiv
Amine Bennouna, Bart Van Parys, Ryan Lucas

Contact

Please reach out if you have an interest in my work.
ryanlu[at]mit.edu
MIT Operations Research Center