Stanford AI: ML with Graphs

Introduction

These days, a growing number of university in the world are putting their courses online on an important variety of subjects, especially regarding Artificial Intelligence (Machine Learning and Deep Learning technologies).

These online courses are ranging from beginner (101), to intermediate and even expert levels.

On this resource we will describe the CS224W course from Stanford University about Machine Learning with Graphs.

Introduction & Description

This course originated in Stanford from Associate Professor Jurij (Jure) Leskovec. It focuses on the computational and algorithmic challenges specific to massive graphs analysis (1).

Complex data can be represented as a graph of relationships between objects. Nowadays these networks are a essential tool for modeling biological, technological and social systems. From the videos and the online material, we are introduced to ML techniques and tools capable of revealing insight on a variety of networks.

Throughout the course, Pr. Leskovec will start with the basics (why do we use Graphs), then continue on Graph Feature methods, embeddings, Deep Learning for Graphs, Graph Neural Network (GNN), and more.

In total the full course is done in over 60 videos (2), 5 colab, exam and a course project. On the official website of the course slides for each course are available as well (1).

Reference

  1. https://web.stanford.edu/class/cs224w/
  2. https://www.youtube.com/playlist?list=PLoROMvodv4rPLKxIpqhjhPgdQy7imNkDn