Fylo›ARCADIA›Graph
Hubs
Association·arcadia

CNNs are special cases of GCNs

Convolutional neural networks can be interpreted as graph convolutional networks applied to regular grids.

Confidence
90%
active

Evidence Quote

“CNNs are a special case of GCNs on regular grids.”

Relationship

Convolutional Neural Networks (CNNs) subtype Graph Convolutional Networks (GCNs)

Arguments

Convolutional Neural Networks (CNNs)subject
Graph Convolutional Networks (GCNs)object

Connections (3)

Flexibility and unification of neural architectures with GNNsInferenceChain
CNNs are special cases of GCNsAssociation
Theoretical and empirical foundations of GNN applicabilityInferenceChain

Evidence

“Reference to Flagel L, Brandvain Y, Schrider DR. (2018), on CNNs effectiveness for population genetic inference.”

Flagel, L. et al. (2018). The Unreasonable Effectiveness of Convolutional Neural Networks in Population Genetic Inference doi:10.1093/molbev/msy224 ↗

“Reference to Cecil RM, Sugden LA. (2023), discussing CNNs for selection inference and the effect of preprocessing.”

Cecil, R.M. & Sugden, L.A. (2023). On convolutional neural networks for selection inference: Revealing the effect of preprocessing on model learning and the capacity to discover novel patterns doi:10.1371/journal.pcbi.1010979 ↗

“Reference to Lecun Y, Bottou L, Bengio Y, Haffner P. (1998), on gradient-based learning for document recognition.”

Lecun, Y. et al. (1998). Gradient-based learning applied to document recognition doi:10.1109/5.726791 ↗

“Reference to Krizhevsky A, Sutskever I, Hinton GE. (2017), regarding deep convolutional networks for ImageNet.”

Krizhevsky, A. et al. (2017). ImageNet classification with deep convolutional neural networks doi:10.1145/3065386 ↗