On Over-Squashing in Message Passing Neural Networks: The Impact of Width, Depth, and Topology.
Francesco Di GiovanniLorenzo GiustiFederico BarberoGiulia LuisePietro LioMichael M. BronsteinPublished in: ICML (2023)
Keyphrases
- message passing
- neural network
- belief propagation
- distributed systems
- probabilistic inference
- inference in graphical models
- factor graphs
- interconnection networks
- approximate inference
- shared memory
- distributed shared memory
- graphical models
- markov random field
- sum product algorithm
- sum product
- depth map
- back propagation
- loopy belief propagation
- artificial neural networks
- ldpc codes
- probabilistic reasoning
- image processing