This is the official code of L2G, Unrolling and Recurrent Unrolling in Learning to Learn Graph Topologies.

Stacy X PU

Last update: Dec 9, 2022

Related tags

Deep Learning L2G-neurips2021

Overview

Learning to Learn Graph Topologies

This is the official code of L2G, Unrolling and Recurrent Unrolling in Learning to Learn Graph Topologies.

Requirement

The code has been tested under:

Python == 3.6.0
PyTorch >= 1.4.0 ｜ CUDA == 10.1

Overview

A quick summary of different folders:

src/models.py contains the source code for the proposed L2G and Unrolling.
src/baselines.py contains the source code for the iterative algorithm PDS and ADMM.
src/utils.py contains utility functions.
src/utils_data.py contains the code for generating synthetic data and graphs.
data/ is a folder for datasets.
log/ contains training logs.
saved_model/ is a folder to store trained models.
saved_results/ is a folder to store testing results.
data_simulation.py contains a code snippet of generating synthetic data and graphs.
main_L2G.py includes the code for training, validating and testing L2G.
main_Unrolling.py includes the code for training, validating and testing Unrolling.

Examples

As there is a requirement on the maximum file size for submissions, we cannot upload all the experimental results and dataset. However, we include all the source code and some of the results as below.

Training and testing L2G on scale-free networks, run:
```
export PYTHONPATH=$PATHONPATH:'pwd' &&
python data_simulation.py &&
python main_L2G.py --graph_type='BA' --n_epochs=100
```
One can find a running log of training and validation loss per epoch at logs/L2G_BA_m20_x20.log. The trained model and test results are automatically saved in saved_model/L2G_BA20_unroll20.pt and saved_results/L2G_BA20_unroll20.pt.

Training and testing Unrolling (ablation study) on scale-free networks, run:

export PYTHONPATH=$PATHONPATH:'pwd' &&
python data_simulation.py &&
python main_Unrolling.py --graph_type='BA' --n_epochs=100

In L2G_WS_m50_x20.ipynb, we show a step-by-step example of training and testing L2G on small-world graphs.

For all the above examples, the results are saved in saved_results/ and the trained models are saved in saved_model/ .

Comments

Performance of L2G_BA20_unrolling

I run your example code in 'README'.md, and I can only get the pred_gmse_mean: 0.96. It is far away from '0.0594' reported in your paper. Is it because of the wrong hyper-parameter? Can I get your all hyper-parameter? Thank your very much

opened by whzyf951620 3

Code for the paper "Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks"

ON-LSTM This repository contains the code used for word-level language model and unsupervised parsing experiments in Ordered Neurons: Integrating Tree

572 Nov 21, 2022

Source Code for our paper: Understand me, if you refer to Aspect Knowledge: Knowledge-aware Gated Recurrent Memory Network

KaGRMN-DSG_ABSA This repository contains the PyTorch source Code for our paper: Understand me, if you refer to Aspect Knowledge: Knowledge-aware Gated

4 May 20, 2022

Differentiable architecture search for convolutional and recurrent networks

Differentiable Architecture Search Code accompanying the paper DARTS: Differentiable Architecture Search Hanxiao Liu, Karen Simonyan, Yiming Yang. arX

3.7k Jan 9, 2023

Pytorch implementation of "Attention-Based Recurrent Neural Network Models for Joint Intent Detection and Slot Filling"

RNN-for-Joint-NLU Pytorch implementation of "Attention-Based Recurrent Neural Network Models for Joint Intent Detection and Slot Filling"

194 Dec 28, 2022

We have implemented shaDow-GNN as a general and powerful pipeline for graph representation learning. For more details, please find our paper titled Deep Graph Neural Networks with Shallow Subgraph Samplers, available on arXiv (https//arxiv.org/abs/2012.01380).

Deep GNN, Shallow Sampling Hanqing Zeng, Muhan Zhang, Yinglong Xia, Ajitesh Srivastava, Andrey Malevich, Rajgopal Kannan, Viktor Prasanna, Long Jin, R

117 Dec 20, 2022

This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning].

CG3 This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning]. R

12 Oct 28, 2022

This is the official code of L2G, Unrolling and Recurrent Unrolling in Learning to Learn Graph Topologies.

Related tags

Overview

Learning to Learn Graph Topologies

Requirement

Overview

Examples

You might also like...

Code for the paper "Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks"

Source Code for our paper: Understand me, if you refer to Aspect Knowledge: Knowledge-aware Gated Recurrent Memory Network

Differentiable architecture search for convolutional and recurrent networks

Pytorch implementation of "Attention-Based Recurrent Neural Network Models for Joint Intent Detection and Slot Filling"

We have implemented shaDow-GNN as a general and powerful pipeline for graph representation learning. For more details, please find our paper titled Deep Graph Neural Networks with Shallow Subgraph Samplers, available on arXiv (https//arxiv.org/abs/2012.01380).

This is the repository for the AAAI 21 paper [Contrastive and Generative Graph Convolutional Networks for Graph-based Semi-Supervised Learning].

The source code of the paper "Understanding Graph Neural Networks from Graph Signal Denoising Perspectives"

Implementation of our paper 'RESA: Recurrent Feature-Shift Aggregator for Lane Detection' in AAAI2021.

OHLC Average Prediction of Apple Inc. Using LSTM Recurrent Neural Network

Comments

Performance of L2G_BA20_unrolling

Owner

Stacy X PU

NEATEST: Evolving Neural Networks Through Augmenting Topologies with Evolution Strategy Training

Flappy bird automation using Neuroevolution of Augmenting Topologies (NEAT) in Python

(ICCV 2021) Official code of "Dressing in Order: Recurrent Person Image Generation for Pose Transfer, Virtual Try-on and Outfit Editing."

This project uses reinforcement learning on stock market and agent tries to learn trading. The goal is to check if the agent can learn to read tape. The project is dedicated to hero in life great Jesse Livermore.

Official implementation for NIPS'17 paper: PredRNN: Recurrent Neural Networks for Predictive Learning Using Spatiotemporal LSTMs.

Official repository for the paper "Going Beyond Linear Transformers with Recurrent Fast Weight Programmers"

Code and datasets for the paper "Combining Events and Frames using Recurrent Asynchronous Multimodal Networks for Monocular Depth Prediction" (RA-L, 2021)

ZSL-KG is a general-purpose zero-shot learning framework with a novel transformer graph convolutional network (TrGCN) to learn class representation from common sense knowledge graphs.

Code for the ECCV2020 paper "A Differentiable Recurrent Surface for Asynchronous Event-Based Data"

the code of the paper: Recurrent Multi-view Alignment Network for Unsupervised Surface Registration (CVPR 2021)