More about HKUST
Efficient Network Embedding for Large-Scale Graph Analysis
Speaker: Dr. Renchi Yang School of Computing National University of Singapore Title: "Efficient Network Embedding for Large-Scale Graph Analysis" Date: Monday, 26 September 2022 Time: 10:00am - 11:00am HKT Zoom link: https://hkust.zoom.us/j/465698645?pwd=aVRaNWs2RHNFcXpnWGlkR05wTTk3UT09 Meeting ID: 465 698 645 Passcode: 20222023 Abstract: In the era of big data, a vital challenge is to efficiently and effectively exploit graph-structured data, such as the World Wide Web, social networks, and biological graphs. In recent years, network embedding emerges as a powerful technique for graph analytics. Network embedding aims to transform nodes in a graph into low-dimensional vectors, which can be fed into off-the-shelf machine learning models used for downstream tasks, and thus, create extensive practical applications in real life. However, most of the existing solutions for network embedding struggle to cope with large graphs (let's say in the order of billions of edges), as they either incur significant computation costs or yield low-quality embeddings on such graphs. In this talk, I will present two effective network embedding approaches that scale to billion-edge graphs with and without node attributes, respectively, using a single commodity machine. The basic idea is to first model the affinity between elements in the graph based on random walks, and then factorize the affinity matrix to derive the embeddings. The main challenges that we address include (i) the choice of the affinity measure and (ii) the reduction of space and time overheads entailed by the construction and factorization of the affinity matrix. Extensive experiments on large graphs demonstrate that our solutions outperform existing methods in terms of both embedding quality and efficiency. ********************* Biography: Dr. Renchi Yang is now a research fellow in the School of Computing, National University of Singapore. Prior to that, he earned his Ph.D. in computer science from Nanyang Technological University in 2021. His research mainly focuses on developing efficient algorithms and systems for large-scale data analysis, with special interests in graph query processing and graph representation learning. He has published more than 12 papers in big data-related top-tier conferences/journals including SIGMOD, VLDB, TODS, KDD, WWW, etc. His research works have received the VLDB 2021 Best Research Paper Award, 2022 ACM SIGMOD Research Highlight Award, and Best Paper Award Nominee in WWW 2022.