At the intersection of computing and mathematics, lie graphs. Graphs are abstract data types meant to leverage graph and hypergraph concepts from mathematics, which are then used in high performance computing (HPC) applications.
Graph computing is a primary research area at the heart of the School of Computational Science and Engineering (CSE), making it one of school’s most identifiable and saught-after areas of study.
Today, we’d like you to meet Xiaojing An, a CSE Ph.D. student with a high performance graph algorithm focus indoors, and a love for all that is outdoors.
Advisor: CSE Chair David A. Bader
Research Focus Areas: High Performance Computing, Graph Algorithms
Hometown: Luoyang, Henan, China
High School: Luoyang No.2 High School
Undergrad Degree: Software Engineering, Zhengzhou University
Current Georgia Tech Degree Program: Ph.D. in Computational Science and Engineering
Tell us about your research:
My research is in high performance graph algorithms, with a focus on improving existing algorithms and their implementations, and developing new algorithms altogether. I am interested in using approximate approaches to reduce algorithmic complexity. I am also interested in graph learning, specifically, using deep learning to solve computationally intensive graph problems. I have worked on graphics processing units (GPUs) and heterogenous systems, along with shared memory systems.
Where is one of your favorite places to hang out in Atlanta?
If you could create something, what would it be?
I would like to create a sublinear approximate graph algorithm. These are algorithms that run in sublinear time, so they do not even read all of the data before returning an answer. They can be useful in the cases when graphs are so large that it takes a lot of time and energy to process, even just iterating through the whole graph and the associated computation is heavy. This seems almost impossible, and while such algorithms cannot give exact answers, in many real cases approximate results are enough.
Last winter break, I wrote a basic review system on top of the Ethereum blockchain system. This review system supports voting and rewards. It also uses blockchains to ensure trustworthiness – in contrast to many existing e-commerce and food reviewing systems – and to support high-quality reviews. I never implemented the front-end to it, but developing the blockchain components and overall structure was a great learning experience!
What are some of your hobbies?
I have a number of different hobbies ranging from rock climbing, hiking, biking, skiing, badminton, handicraft to cooking, watching movies, and more!
What is a favorite memory from your years at Georgia Tech?
My favorite memory during my time at Georgia Tech is when I went to Supercomputing 2017 in Denver. I was able to establish connections with researchers, I gained a broader understanding of the field of HPC, and I was exposed to some very interesting presentations. Denver is also a beautiful city! Some labmates and I hiked in Rocky Mountain National Park while it snowed and had a great time.