🔴🟨 Edge#203: What are Graph Recurrent Neural Networks?
+ what GNNs on Dynamic Graphs; and the exploration of DeepMind’s Jraph, a GNN Library for JAX.
In this issue:
we explain what Graph Recurrent Neural Networks are;
we discuss GNNs on Dynamic Graphs;
we explore DeepMind’s Jraph, a GNN Library for JAX.
Enjoy the learning!
💡 ML Concept of the Day: What are Graph Recurrent Neural Networks?
Recurrent neural networks (RNNs) are one of the most popular architectures in modern deep learning. In GNN theory, the equivalent to RNNs is represented by an architecture known as graph recurrent neural networks (GRNNs). As its name indicates, the core idea of GRNN is to generalize RNN principles used for sequential data processing to process graph data. Just like RNNs learn dependencies over sequential datasets, GRNNs can do something similar for graph-structured data. This problem can be particularly complicated in graphs as nodes can have an arbitrary number of relationships.
From an architecture standpoint, a GRNN is a GNN that uses
Keep reading with a 7-day free trial
Subscribe to TheSequence to keep reading this post and get 7 days of free access to the full post archives.