Stanford CS224W: ML with Graphs | 2021 | Lecture 4.4 – Matrix Factorization and Node Embeddings

For more information about Stanford’s Artificial Intelligence professional and graduate programs, visit: https://stanford.io/2XRqWKq

Jure Leskovec
Computer Science, PhD

In the last part of this lecture, we switch gears a bit and discuss matrix factorization method for generating node embeddings. Specifically, we discuss how previously mentioned methods for learning node embeddings connect to methods in matrix factorization. With these intuitions, we then present 3 limitations with embedding methods based on matrix factorization and random walks. Coming up in the next lectures, we will introduce specific solutions to these limitations: Deep Representation Learning and Graph Neural Networks.

To follow along with the course schedule and syllabus, visit:
http://web.stanford.edu/class/cs224w/