Me calling Yi Lai, golden gate bridge.

Angxiu Ni (pronounces as "ang sh-you knee")

Assistant professor, Department of Math, UC Irvine.
E-mail: angxiun at uci dot edu

Recent

I generalized the kernel-differentiation method to an ergodic version and a foliated version.
The continuous-time equivariant divergence formula (with Yao Tong) is acpublished at JSP.
The adjoint shadowing lemma for discrete and contiuous-time hyperbolic systems is accepted at Nonlinearity.

Brief description of Research

I differentiate chaos, or to compute the parameter-derivatives of averaged observable of chaos. I gave the pointwisely defined linear response formula for hyperbolic system, making Monte-Carlo sampling (when dynamics is given, MC means to sample by a long orbit) possible. More specifically, the tools we developed
1. Fast tangent response formula, equivariant divergence formula, and algorithms.
2. Ergodic and foliated kernel differentiation (or likelihood ratio) method.
3. Adjoint theories and algorithms, backpropagation under gradient explosion.
4. Non-intrusive shadowing algorithms.
I am also interested in the numerical interaction with chaos in all fields, such as fluids, geophysics, inference, data assimilation, and machine learning.

Publications and Talks


Teaching, Winter 2025

Linear algebra 2, UC Irvine.
Machine learning, UC Irvine.

Misc.

Curriculum vitae
Prvious teaching
I got my PhD from UC Berkeley math, did a postdoc at PKU BICMR, then worked as Assistant professor at and YMSC in Tsinghua University, where I also served as Manager of undergrad affairs in Qiu Zhen College.
I am or was mentored by John Strain, Pingwen Zhang, Mark Pollicott, Jack Xin, Qing Nie, Long Chen.
Previous Teaching