After clicking “Watch Now” you will be prompted to login or join.
Recommendation Systems Using Distributed Graph Convolutional Networks (GCN) on GPUs (Presented by Amazon Web Services)
Cyrus Vahid, Amazon Web Services
GTC 2020
Deep learning so far has been mostly applied to simple and structured data, such as images, or sequences, such as time-series and language. Most of the information in the world is, however, non-Euclidean and complex and can be presented as graph. Molecules, social network relatedness, and relationships between products and consumers are among the more popular examples. We'll represent an intuitive theory of Graph Convolutional Networks and provide a walkthrough on an implementation for recommender systems using Deep Graph Library and Apache MXNet.