Graph Neural Networks in Flux.


Im starting my master course (Brazil- UFPR) in Bioinformatics. My background is mostly in IT and I´ve been studing Neural Nets(CNN, RNN…) to my upmost abilities.

It seems that Graphs Neural Networks and/or Graphs Convolutional Networks are sort of new, and there´s little information about them.

Since Graphs are widely used in Bioinfo( Molecular structures, proteins structs, even Dna and more) I think it´s a good proposal for a paper.

Flux is amazing, and I think it would be a good framework to build GNN on.

If anyone could point me into a direction or even talk about it, I thank you in advance.


There’s been talk of implementing graph neural networks in the Flux GitHub repo but I’m not aware of a formal implementation as of yet. My opinion while learning these subjects is that they are both really simple taken separately. If you just practice using Flux for your basic neural networks to gain some intuition it probably wouldn’t be hard to implement something novel using something like LightGraphs


See for that aforementioned discussion.

Just a heads up - @yuehhua started the project called GeometricFlux.jl and it’s planned to be included in the FluxML project itself. See the announcement on this forum for more information.

To notify the non-overlapping people, I double-post the same words.

Thanks to @XVilka for promoting GeometricFlux.jl. I have implemented some frequently used geometric deep learning/graph neural network layers and the abstraction of message-passing scheme, even the graph network (GN) block abstraction. PRs are welcome. Documentation and tutorials are coming.


We also have GitHub - CarloLucibello/GraphNeuralNetworks.jl: Graph Neural Networks in Julia now.
In a few months, I will move this project to the FluxML org.