Learning Theory Can (Sometimes) Explain Generalisation in Graph Neural Networks.

Pascal Mattia Esser Leena C. Vankadara Debarghya Ghoshdastidar

Published in: NeurIPS (2021)

Keyphrases

learning theory
neural network
computational learning theory
instructional design
widely applicable
graph theory
concept classes
automata theory
back propagation
pac learning
theoretical computer science
objective function
neural network model
learning theories
generalization bounds