Beyond message passing: A physics-inspired paradigm for graph neural networks

https://thegradient.pub/graph-neural-networks-beyond-message-passing-and-weisfeiler-lehman/

by andreyk • 4 years ago

94 16 4 years ago

thegradient.pub

melony• 4 years ago

Are implementations of belief propagation considered message passing GNNs?

andreyk2• 4 years ago

Pretty sure that's the case, yeah

phonebucket• 4 years ago

Can anyone recommend any arXiv/paper links on the subject for someone with reasonable prerequisties (e.g. neural ODEs, physics informed neural networks and message passing)? The number of references in the article is a bit of an overload! Looks like fascinating field.

andreyk2• 4 years ago

A survey paper is usually a good way to go, such as A comprehensive survey on graph neural networks (https://arxiv.org/abs/1901.00596) or Graph Neural Networks: A Review of Methods and Applications (https://arxiv.org/abs/1812.08434)

albertzeyer• 4 years ago

Can someone explain the downvotes here? If you think these are bad papers, maybe recommend some better ones? Or what is wrong with this post?

ssivark• 4 years ago

From a quick glance, the blog post seems to be based on the following paper involving the same author: https://arxiv.org/abs/2106.10934

The point of the post is how graph NNs are limited in capability, so I’m not sure graph NN surveys are the best references (sibling comment) for the main punchline, though they might certainly contain many important details.

sandGorgon3• 4 years ago

anyone running graph neural networks in production ? what framework do you use ?

dil8• 4 years ago

Check out dgl (https://github.com/dmlc/dgl). A lot of papers and algorithms are implemented in the examples section.

amelius• 4 years ago

Isn't this used for predicting protein folding/interactions?

sandGorgon3• 4 years ago

that's like Google's experimental work. I mean Google & Facebook using something is kinda moot.

Is anyone non-FAANG using it ?

hasmanean• 4 years ago

So when will we have message-passing processor architectures again?

wnoise• 4 years ago

They do actually exist. Caltech has (had?) a group doing asynchronous microprocessors, which were most naturally expressed as channels between computation units. (They had a subset of CSP they called CHP -- communicating hardware processes that they used for a lot of the high-level designs).

As usual with nice things, the market didn't actually want it. Though Intel did buy a startup they spun out and used it mostly for switch fabric purposes.

pizza• 4 years ago

This is good stuff. One really cool use of sheaf theory is being able to predict race conditions between 2 states in switching electrical circuits by showing H_1 ("line-like", if H_0 is "point-like") cohomology isomorphic (? not sure if that's the right term) to Z_2: https://www.drmichaelrobinson.net/sheaftutorial/20150826_tut... (slide 29)

Beyond message passing: A physics-inspired paradigm for graph neural networks

Related Stories