Login

Tutorials

Tutorials 1

Machine learning on Graphs

 

 

Prof. Chanyoung Park

KAIST, Korea

 

Abstract

Graphs are data structures that express the connection relationships between individuals, and are widely used to express various phenomena in real life. Representative examples include user's social networks, knowledge graphs, molecular structure graphs, protein-protein interaction graphs, and gene graphs. In order to improve the performance of graph-based machine learning models, it is essential to learn the representation of nodes and edges in consideration of the structure of the graph. To this end, with the recent development of deep learning technology, machine learning techniques for graph analysis are in the spotlight.

This tutorial aims to provide an introduction to the basics of recent advances in machine learning on graphs. More specifically, I will mainly discuss about recent machine learning techniques to learn representations of nodes in various types of graph, including homogeneous, multi-aspect, attributed, and heterogeneous graph. Moreover, I will introduce interesting applications of graph machine learning. This tutorial is self-contained and designed for introductory and intermediate audiences. No special prerequisite knowledge is required to attend this tutorial.

 

Bio

Chanyoung Park is an assistant professor in the Dept. of Industrial and Systems Engineering and Graduate School of AI at KAIST. He received his Ph.D from the Dept. of Computer Science and Engineering at POSTECH in 2019. Before joining KAIST, he was a postdoctoral research fellow in the Dept. of Computer Science at University of Illinois at Urbana-Champaign. His research focuses on developing graph-based machine learning methodologies and their applications, including graph representation learning, user behavior analysis, graph neural networks for chemistry/bioinformatics, and etc.

Tutorials 2

Multi-resolution Graph Analysis for Graphical Model Selection and Graph Classification

 

 

Prof. Won Hwa Kim

POSTECH, Korea

 

Abstract

 

1. Topic of the tutorial

  • Title: Multi-resolution Graph Analysis for Graphical Model Selection and Graph Classification
  • Target audience: Whoever interested in graph data analysis (in Neuroimaging)
  • Prerequisite: basic knowledge in Linear Algebra

 

2. Outline

1. Introduction to Multi-resolution (Basics of signal transform, spectral graph theory, graph signal processing)

2. Latent Graphical Model Selection via Multi-resolution Analysis

  • Latent variable graphical model selection via convex optimization, Chandraskaran et al., The Annals of Statistics, 2012
  • Latent Variable Graphical Model Selection using Harmonic Analysis: Applications to the Human Connectome Project (HCP), Kim et al., Computer Vision and Pattern Recognition (CVPR), 2017

3. Learning Multi-resolution Representation for Human Connectome Classification

  • Learning Multi-resolution Graph Edge Embedding for Discovering Brain Network Dysfunction in Neurological Disorders, Ma et al., Information Processing in Medical Imaging (IPMI), 2021
  • Enriching Statistical Inferences on Brain Connectivity for Alzheimer's Disease Analysis via Latent Space Graph Embedding, Ma et al., International Symposium on Biomedical Imaging (ISBI), 2020

 

Bio

1. Name: Won Hwa Kim

2. Affiliation:

  • Graduate School of AI / Computer Science and Engineering at POSTECH
  • Computer Science and Engineering at the University of Texas at Arlington

3. Email: wonhwa@postech.ac.kr

4. Bio: Dr. Won Hwa Kim is an Assistant Professor in Graduate School of AI / CSE at POSTECH, CSE at the University of Texas at Arlington (currently on leave). He obtained his Ph.D in Computer Sciences from University of Wisconsin - Madison in 2017, M.S. in Robotics from KAIST in 2010 and B.S. in Information and  Communication Engineering from Sungkyunkwan University in 2008. Prior to joining academia, he worked as a researcher in Data Science team at NEC Labs., America in 2017. He focuses on interdisciplinary research that crosses cores in Computer Vision, Machine Learning and Neuroscience, developing novel methods for analyses of data in non-Euclidean spaces and their applications in Image Analysis including Neuroimaging. His research has been supported by several US federal agencies such as USDOT, NIH and NSF, and he is an NSF CRII awardee.

5. Background in the tutorial area: Dr. Kim’s research work on graph analysis has been published in top-tier AI conferences such as NIPS, CVPR, ICCV, ECCV, MICCAI, ISBI as well as in high impact journals such as TPAMI, NeuroImage, NeuroImage: Clinical and Brain Connectivity.

 

 

 

ⓒ Copyright 2021 KIISE – All Rights Reserved.

[KIISE] Korean Institute of Information Scientists and Engineers

#401 Meorijae Bldg., 76, Bangbae-ro, Seocho-gu, Seoul 06704, Korea

Email: mkim@kiise.or.kr | Fax: +82-2-521-1352 | Homepage: www.kiise.or.kr

Business Registration Number: 114-82-03170

SCMember-board