Overview
Explore a comprehensive 40-minute video walkthrough of a Graph Attention Network (GAT) project implementation. Dive into the Cora dataset, learn about a highly-optimized GAT implementation, and discover other exciting deep learning projects. Follow along as the instructor covers node degree statistics, entropy histograms, t-SNE plots, and graph drawing layouts. Gain insights into the Cora dataset, feature vectors, labels, and edge index construction. Understand the implementation through a toy example, explore lifting techniques, and grasp neighborhood-aware softmax and aggregation methods. Perfect for those looking to deepen their understanding of GAT and its practical applications in deep learning.
Syllabus
Intro to GAT project
My other deep learning projects
README walkthrough
Node degree statistics
Entropy histograms
t-SNE plots
Graph drawing layout
Jupyter walkthrough
Understanding Cora dataset
Feature vectors and labels
Building the edge index
Toy example understanding the implementation
Lifting
Neighborhood aware softmax and aggregate
Outro, exciting deep learning projects
Taught by
Aleksa Gordić - The AI Epiphany