PIMS PDF Hui Huang: On Big Leaps, Dynamical Systems and Partial Differential Equations.
By Ruth A. Situma, PIMS Program & Communications Manager
As we get back to in-person learning and teaching, we connected with our PIMS PDF Hui Huang who is at the University of Calgary. Hui got to Calgary at the beginning of 2021 and in his words loves “ the cowboy city where you can see the Rocky Mountains”. Since arriving, he has been trying to achieve a balance in work and other activities. In Calgary, he lives close to Prince’s Island Park where he can take walks along the river and people-watch. Hui was a regular at the gym before the pandemic and he is hoping he can pick that up again soon. “On weekends I enjoy watching old movies, especially Hitchcock mystery movies. I also read novels and have been enjoying the free concerts being held in Calgary this summer,” says Hui.
You have made a number of great leaps, moved from a small town to big city, obtained a doctorate from Tsinghua University and you are now starting a third postdoctoral term. Your academic journey is one of resilience, tell us about it?
I grew up in a small mountain village where I had to walk an hour to get to school. My village is in the province of Fujian on the southern coast of China. We only had about 10 households in the village and all of us children played in the fields together. For Middle School and High School, I had to move to the larger town nearby and visited home on weekends and holidays. Luckily I was accepted to a university in Beijing in 2008. That was my first time taking a train and the trip took about 24 hours. It opened my eyes to see what a big city looks like, especially because Beijing was hosting the Olympic games that year. My next big move was to Duke University in Durham, North Carolina where I spent 2 years as a joint Ph.D. student. My thesis was completed under the consultations of Prof. Jian-Guo Liu and Prof. Huai-Yu Jian who introduced me to the academic world. This was my first time overseas and it was an interesting project to understand the culture there.
Once I got my doctoral degree from Tsinghua University in 2017, I embarked on a series of postdoc positions in interesting places. First I was at Simon Fraser University in Vancouver, BC as an Alan Mekler postdoctoral fellow, where I worked with Prof. Razvan Fetecau and Prof. Weiran Sun. This was my first time in Canada and I realized what a great place it is. I loved having the ocean on one side and the snow-capped mountains on the other. Then I went to Germany where I worked at the Technical University of Munich with Prof. Massimo Fornasier. This also expanded my horizons and I met many wonderful mathematicians there.
You have done a lot of work on dynamical systems, a very broad area in mathematics. How does this research tie in with your collaborations with Prof. Jinniao Qiu at the University of Calgary?
My research focuses on the analysis of complex dynamical systems and their related kinetic equations. This ends up touching on a lot of interesting subjects, such as partial differential equations, probability theory, stochastic calculus, numerical analysis, optimization, and control theory. In particular, I spend a lot of time working on the derivation of aggregation-diffusion type differential equations from (stochastic) interacting particle systems, which may arise from mathematical modeling of swarming and self-collective behavior of complex biological systems. This is related to the concept of mean-field limit and the theory of propagation of chaos.
Recently in Munich, I have been working on nonlinear dynamical systems in machine learning. Thanks to Massimo, my supervisor there, for introducing me to this field. More specifically we are interested in the particle models that belong to the class of optimization methods called metaheuristics. They can provide empirically robust solutions to tackle hard optimization problems with fast algorithms. This has led me to some of my latest research contributions, including parameter identification, energy inference, consensus-based optimization, and particle swarm optimization for non-convex functions, which has vast potential for application in signal processing and machine learning.
When I moved to Calgary I brought these projects along too. I have been working on them further with my current supervisor Prof. Jinniao Qiu. At the same time, we are starting a new project on solving partial differential equations using deep learning. Of course, I still keep in close touch with the team in Munich too.
The last few seasons have been “strange”, and many of us are moved to look at life through a different lens. What are you most thankful for despite the Covid lockdown?
First of all, I’m so glad that none of my friends or family has caught Covid-19. I was especially worried in the beginning when it started in China that it might affect my family, but luckily they have been okay thanks to the government’s strict lockdown. I’m grateful to my supervisors Jinniao and Massimo for supporting me through this stressful time. Also, I appreciate the danger that our front-line workers have had to risk to keep society running.
Hui Huang will be speaking at the PIMS Emergent Research Seminar Series, September 15, 2021 at 9:30AM Pacific. Details on his talk (Large systems of interacting particles and their application in optimization) can be found here.
