An attempt at demystifying graph deep learning
ericmjl.github.io
An attempt at demystifying graph deep learning
1–10 of 22 posts
Re: An attempt at demystifying graph deep learning
#2If you understand transformers, you understand the basics behind any neural "graph network" method. They are exceedingly simple
Re: An attempt at demystifying graph deep learning
#3If you understand transformers, you understand the basics behind any neural "graph network" method. They are exceedingly simple
Oh! Is that what I've been missing for two years?
Well, when you phrase it like that... Hmm. Yup, I'm an idiot.
<3
Re: An attempt at demystifying graph deep learning
#4In case you're interested in learning about graph deep learning, and are familiar with standard DL, I strongly recommend these two very good, recent books (freely available):
[1] Graph Representation Learning - William L. Hamilton https://www.cs.mcgill.ca/~wlh/grl_book/
[2] Deep Learning on Graphs - Yao Ma, Jilian Tang https://cse.msu.edu/~mayao4/dlg_book/
Re: An attempt at demystifying graph deep learning
#5What's the current killer app of GNNs? By that I mean, in which tasks is using a GNN-based approach obviously better?
Conceptually, just by framing the graph as a mxm matrix I assume the most obvious comparison would be against some sort of dimensionality reduction for node embedding. Should I see graph(node) embedding as an alternative to e.g. PCA, LDA, t-SNE and UMAP and variants?
Re: An attempt at demystifying graph deep learning
#6If you understand transformers, you understand the basics behind any neural "graph network" method. They are exceedingly simple
A monad is just a monoid in the category of endofunctors
Re: An attempt at demystifying graph deep learning
#7A good survey paper is here (Jan 2021) by Chami, Abu-El-Haifa, Perozzi, Re and Murphy
[2005.03675] Machine Learning on Graphs: A Model and Comprehensive Taxonomy https://arxiv.org/abs/2005.03675
Re: An attempt at demystifying graph deep learning
#8Tangential — Curious how the author made those diagrams
Re: An attempt at demystifying graph deep learning
#9Tangential — Curious how the author made those diagrams
Looks like Notability colours/lines/shapes to me!
Re: An attempt at demystifying graph deep learning
#10What's the current killer app of GNNs? By that I mean, in which tasks is using a GNN-based approach obviously better? Conceptually, just by framing the graph as a mxm matrix I assume the most obvious comparison would be against some sort of dimensionality reduction for node embedding. Should I see graph(node) embedding as an alternative to e.g. PCA, LDA, t-SNE and UMAP and variants?
It’s pretty clearly the current best way to do deep learning on molecules in chemistry. Transformers do really well in predicting reaction outcomes but GCNNs are still better when it comes to “is this a good molecule according to this label” type questions. That said, we have added attention layers to GCNNs so I think the lines get blurry.