tech-srl / bottleneck Goto Github PK
View Code? Open in Web Editor NEWCode for the paper: "On the Bottleneck of Graph Neural Networks and Its Practical Implications"
License: MIT License
Code for the paper: "On the Bottleneck of Graph Neural Networks and Its Practical Implications"
License: MIT License
Hi Uri,
Thank you for this amazing and insightful work. I have the following two questions about the experiment. Hope I can get your help.
(1) Did you run experiments for GNNs+FA on the NeighborsMatch dataset? If yes, could you share the results?
(2) I didn't find the results for GIN on the NeighborsMatch dataset when r=7/8. Is there any reason about this?
Best,
Meng
i am confused for the input data of GIN model code part:
x, edge_index, batch = data.x, data.edge_index, data.batch
could you explain the three ones?
Hi Uri,
I hope you're doing well. Your paper is really nice, congrats.
I just have a small question. If I correctly understood your code, you're generating several trees and then you stack them to train in a batch fashion. What I don't understand is why you add self loops to all nodes. The illustrations in the paper don't show any self loops. I guess the only difference is that you'll be using the own embedding of each node when performing message passing, isn't it? Or is there something I'm missing?
Thank you very much for releasing the code :)
Hi,
This is an interesting paper. may find some fundamental issues about GNN.
I have a question about the Tree-NeighborsMatch problem. In my understanding, 1-layer GNN can pretty much solve it, because what the model has to do is:
In this case, it doesn't have to message pass all the information from leaves nodes to target nodes to find out the most similar one. It doesn't need to "match", it just needs to learn 2 --> "C"
Hi, Uri Alon, I love this work. I think your paper is splendid and innovative to follow. And I want to know whether and where this paper is published. I can't wait to cite your paper as my support.
Hello,
Thanks for this great repo. I cannot reproduce the exact results with the default configurations as follows:
python run-gcn-2-8.py
python run-gat-2-8.py
python run-ggnn-2-8.py
python run-gin-2-8.py
The results produced by these scripts as as follows:
2 | 3 | 4 | 5 | |
---|---|---|---|---|
GGNN | 1 | 0.99 | 0.93 | 0.49 |
GAT | 1 | 1 | 0.63 | * |
GIN | 1 | 0.99 | * | * |
GCN | 1 | 1 | 0.58 | * |
Marker * denotes the experiments are still running.
May I ask the hyper-parameters in the default configurations are the same as the paper used?
Thanks.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.