Explore graph neural networks (GNNs) in this 21-minute conference talk from YOW! 2018. Delve into the architecture and applications of GNNs, a variant of deep neural networks designed to model data represented as generic graphs. Learn about graph representation, including graph of graphs (GoGs), and how different data types can be represented using graphs. Discover the architecture of deep graph neural networks and their learning algorithms. Examine practical applications of GoGs and GNNs, such as document classification, web spam detection, and human action recognition in video. Gain insights into how GNNs differ from convolutional neural networks and their potential in advancing artificial intelligence across various domains.
Graph Neural Networks - Algorithm & Applications
Overview
Syllabus
Introduction
Definitions
Data representations
Web document categorization
Graph of graphs
Encoding and Output Network
Learning Process
Learning Process Summary
Applications
Action Recognition
Conclusion
Taught by
GOTO Conferences
