I’m Linkai Ma, a 3rd year Ph.D. student in Computer Science at Purdue University, advised by Petros Drineas. Before Purdue, I completed my B.A./M.S. in Mathematics at NYU’s Courant Institute of Mathematical Sciences, where I was fortunate to work with Professor Dimitris Giannakis for mathematical modeling of sea ice and Professor Olivier Pauluis for autoregressive DMD model of atmospheric annular mode. I also completed my master thesis under the supervision of Professor Jonathan Weare.
My current research sits at the intersection of numerical linear algebra, randomized algorithms, and theoretical machine learning. I am especially interested in the regularization effects of stochastic rounding/quantization on linear systems and more complex models such as neural networks. My broader goal is to connect rigorous analysis (perturbation theory, concentration, random matrix tools) with scalable, reliable numerical methods for modern ML systems.
Projects
- Stochastic Rounding Implicitly Regularizes Tall-and-Thin Matrices When/how stochastic rounding regularizes the smallest singular value of a matrix.
- A Note on the Stability of the Sherman-Morrison-Woodbury How does matrix inversion errors propagate through SMW.
Teaching
At Purdue, I’ve been a teaching assistant for both undergraduate and graduate level numerical analysis courses. At NYU, I was a TA for calculus I.
A bit more about me
Outside research, I’m really into music, traveling, and basketball.
Contact
- Email: ma856@purdue.edu