Mars (Liyao) Gao
Postdoctoral Scholar
Stanford University
Email: marsgao [at] stanford [dot] edu
GitHub |
Google Scholar
About me
I am a postdoctoral scholar at Stanford University and a Stanford AI Lab (SAIL) Fellow, advised by Professor Emily B. Fox. I received my Ph.D. from the Paul G. Allen School of Computer Science & Engineering at the University of Washington, where I was advised by Professor J. Nathan Kutz.
My research centers on AI for science (AI4S): interpretable and adaptive learning frameworks that recover governing structure from complex scientific data, quantify the uncertainty behind it, and keep working when physical conditions shift at test time. Recent work includes sparse and Bayesian model discovery methods built on SINDy and Koopman theory, shallow recurrent decoders for sensing high-dimensional spatiotemporal fields, and data assimilation frameworks for sparse, noisy measurements — applied across biology, physiology, physics, astronomy, and power systems.
Working across these domains has pointed me toward a broader question: how AI systems can participate directly in the scientific discovery loop — forming hypotheses, running and interpreting experiments, and revising models — rather than serving only as a modeling tool. My long-term goal is to build AI helpers that reason over complex scientific data alongside human researchers and produce insights they can trust. More on my research directions →
News
Selected publications
 |
Sparse identification of nonlinear dynamics and Koopman operators with Shallow Recurrent Decoder Networks. |
 |
SENDAI: A Hierarchical Sparse-measurement, EfficieNt Data AssImilation
Framework. |
 |
Bayesian autoencoders for data-driven discovery of coordinates, governing equations and fundamental constants. |
 |
Convergence of uncertainty estimates in ensemble and Bayesian sparse model discovery. |
 |
Deformation Robust Roto-Scale-Translation Equivariant CNNs. |
 |
On Optimal Early Stopping: Over-informative versus Under-informative Parametrization. |