Together with Swedish professor, ‘simplify’ many problems with weapon of Graph Neural Network (GNN)

At 4 pm on February 23, FSOFT AI Center will hold a knowledge-sharing webinar titled “Graph neural network: A powerful ‘axe’ to ‘chop’ down many learning problems” with the participation of Professor Xuan Son Vu from Umeå University (Sweden). Webinar promises to bring valuable knowledge about graph neural networks and their applications in practice.

>> Registration link: https://forms.office.com/r/0ARLPdLRxk

GNNs learn data that is structured using deep learning, specifically a combination of End-to-end Learning and Inductive Reasoning. This is a promising research direction that can address causal reasoning and interpretability, which are the main limitations of traditional Deep Learning.

GNNs can be classified into 3 types including (1) Wandering model (e.g. DeepWalk), (2) Message passing (e.g. GCN), (3) Knowledge graph model (e.g. TransE).

In this webinar, FSOFTer will share specifically the cases in which GNN is used, thereby in-depth discussion of recent research works on learning privacy protection graphs and modular graph networks for multi-label classification tasks, and at the same time relate these concepts to real-life applications.

The event was led by Dr. Xuan Son Vu – Member of the Institute of Electrical and Electronics Engineers (IEEE). He graduated with a Bachelor’s degree in Information Systems from the University of Technology, Hanoi National University in 2011 and a Master’s degree from Kyungpook National University in 2014. In 2020, he received his PhD in Computer Science. from Umeå University, Sweden with a focus on big data privacy-preserving machine learning. Before joining Umeå University, from 2015 to 2016 he was a full member of UKPLab, TU Darmstadt, Germany (led by Professor Iryna Gurevych, future President of the Society for Computational Linguistics – ACL in 2023).

He is currently doing a Postdoctoral Fellow at Umeå University, Sweden, where he focuses his research on Robust Machine Learning. His research interests focus on obtaining knowledge from multimodal data and using structured knowledge to power downstream applications.

Professor Xuan-Son Vu’s awards and honors include 3rd place for the best paper award at CICLING 2019, best student paper award and most inquisitive intellectual award at CICLING 2018. He also received the best article award at the 40th Conference of Korea Information Processing Society (KIPS) 2014.