Graph Neural Networks

KGCN

Knowledge Graph Convolutional Network (github). #todo

Backlinks

• [[graph-network-as-arbitrary-inductive-bias]]
  • The architecture of a neural network imposes some kind of structure that lends itself to particular types of problem (CNN, RNN). Thus, you can think of this as some form of inductive bias. An interesting view of [[graph-neural-networks]] is that essentially these provide arbitrary inductive bias, since the goal is to learn the architecture?
• [[bitter-lesson]]
  • What is pretty clear is that domain-knowledge injection is not scalable (think expert systems back in the day). What’s better are general-purpose methodologies, and the more general-purpose, the better. However, this would suggest that something like CNNs are actually suboptimal, since convolutions are definitely highly specific to image classification. But I think that’s part of the allure of [[graph-neural-networks]].