Why is "Hidden State" at Node in Graph Neural Network Considered "Compressed" Representation?
I am reading article on Graph Neural Network (GNN) and it is mentioned:
The memory stores the states of all the nodes, acting as a compressed representation of the node’s past interactions.
Why is the hidden state or node's memory considered compressed representation of the node's interaction? Isnt compressed representation is what auto-encoders yield?
Topic graph-neural-network
Category Data Science