susheels / adgcl Goto Github PK
View Code? Open in Web Editor NEWAdversarial Graph Augmentation to Improve Graph Contrastive Learning
License: MIT License
Adversarial Graph Augmentation to Improve Graph Contrastive Learning
License: MIT License
Hi, thanks for your great work!
I have a question about the detail of NAD-GCL.
In 5.1 of the paper it writes:
NAD-GCL drops the edges of a graph uniformly at random. We consider NAD-GCL-FIX and NAD-GCL-OPT with different edge drop ratios. NAD-GCL-GCL adopts the edge drop ratio of AD-GCL-FIX at the saddle point of the optimization (Eq.8) while NAD-GCL-OPT optimally tunes the edge drop ratio over the validation datasets to match AD-GCL-OPT.
I didn't quite understand how to define the edge drop ratio in NAD-GCL.
What's the difference between NAD-GCL and GraphCL which using EdgePert?
Thank you!
Hi @susheels
Thanks for the great work. I tried to reproduce the transfer learning results of AD-GCL. Concretely, I pretrained the model on the ZINC-2M for 100 epochs, and fine-tuned it on the downstream tasks. However, the reproduced results are lower than ones in the paper. Could you pls help me with it? Thanks!
Hi, thanks for your excellent work!
I find that you evaluate adgcl mainly on graph-level task with multiple datasets. And I wonder whether adgcl can be applied on a single graph dataset like Cora or Citeseer for node classification?
Hi @susheels
Thanks for your great work. I have a minor request that could you pls release the code of baselines (e.g., GraphCL) for OGBG. I think it's a bit difficult to adapt the test_minimax_ogbg.py
directly. It's really helpful if you could release them. Thanks a lot!
Hi! Could you please your pretrained model files of transfer learning(bio and chem dataset)? Thus I can use it for finetuning and better approximate the results of transfer learning in your paper.
Hi!
I'm trying to reproduce your AD-GCL results on TU-dataset (unsupervised learning). I could achieve nearly the same results as your paper reported. However, when I split the training process into two steps (AD_GCL for latent vector generation and using latent vector for linear classification (linear SVC). The result is bad enough (training : 65%, val and test: ~50%)). I wonder what is the difference between these two training strategies.
Thanks a lot
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ 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. ๐๐๐
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 โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.