Expressive Power of Invariant and Equivaraint Graph Neural Networks (ICLR 2021)

Overview

Expressive Power of Invariant and Equivaraint Graph Neural Networks

In this repository, we show how to use powerful GNN (2-FGNN) to solve a graph alignment problem. This code was used to derive the practical results in the following paper:

Waiss Azizian, Marc Lelarge. Expressive Power of Invariant and Equivariant Graph Neural Networks, ICLR 2021.

arXiv OpenReview

Problem: alignment of graphs

The graph isomorphism problem is the computational problem of determining whether two finite graphs are isomorphic. Here we consider a noisy version of this problem: the two graphs below are noisy versions of a parent graph. There is no strict isomorphism between them. Can we still match the vertices of graph 1 with the corresponding vertices of graph 2?

graph 1 graph 2

With our GNN, we obtain the following results: green vertices are well paired vertices and red vertices are errors. Both graphs are now represented using the layout from the right above but the color of the vertices are the same on both sides. At inference, our GNN builds node embedding for the vertices of graphs 1 and 2. Finally a node of graph 1 is matched to its most similar node of graph 2 in this embedding space.

graph 1 graph 2

Below, on the left, we plot the errors made by our GNN: errors made on red vertices are represented by links corresponding to a wrong matching or cycle; on the right, we superpose the two graphs: green edges are in both graphs (they correspond to the parent graph), orange edges are in graph 1 only and blue edges are in graph 2 only. We clearly see the impact of the noisy edges (orange and blue) as each red vertex (corresponding to an error) is connected to such edges (except the isolated red vertex).

Wrong matchings/cycles Superposing the 2 graphs

To measure the performance of our GNN, instead of looking at vertices, we can look at edges. On the left below, we see that our GNN recovers most of the green edges present in graphs 1 and 2 (edges from the parent graph). On the right, mismatched edges correspond mostly to noisy (orange and blue) edges (present in only one of the graphs 1 or 2).

Matched edges Mismatched edges

Training GNN for the graph alignment problem

For the training of our GNN, we generate synthetic datasets as follows: first sample the parent graph and then add edges to construct graphs 1 and 2. We obtain a dataset made of pairs of graphs for which we know the true matching of vertices. We then use a siamese encoder as shown below where the same GNN (i.e. shared weights) is used for both graphs. The node embeddings constructed for each graph are then used to predict the corresponding permutation index by taking the outer product and a softmax along each row. The GNN is trained with a standard cross-entropy loss. At inference, we can add a LAP solver to get a permutation from the matrix .

Various architectures can be used for the GNN and we find that FGNN (first introduced by Maron et al. in Provably Powerful Graph Networks NeurIPS 2019) are best performing for our task. In our paper Expressive Power of Invariant and Equivariant Graph Neural Networks, we substantiate these empirical findings by proving that FGNN has a better power of approximation among all equivariant architectures working with tensors of order 2 presented so far (this includes message passing GNN or linear GNN).

Results

Each line corresponds to a model trained at a given noise level and shows its accuracy across all noise levels. We see that pretrained models generalize very well at noise levels unseen during the training.

We provide a simple notebook to reproduce this result for the pretrained model released with this repository (to run the notebook create a ipykernel with name gnn and with the required dependencies as described below).

We refer to our paper for comparisons with other algorithms (message passing GNN, spectral or SDP algorithms).

To cite our paper:

@inproceedings{azizian2020characterizing,
  title={Expressive power of invariant and equivariant graph neural networks},
  author={Azizian, Wa{\"\i}ss and Lelarge, Marc},
  booktitle={International Conference on Learning Representations},
  year={2021},
  url={https://openreview.net/forum?id=lxHgXYN4bwl}
}

Overview of the code

Project structure

.
├── loaders
|   └── dataset selector
|   └── data_generator.py # generating random graphs
|   └── test_data_generator.py
|   └── siamese_loader.py # loading pairs 
├── models
|   └── architecture selector
|   └── layers.py # equivariant block
|   └── base_model.py # powerful GNN Graph -> Graph
|   └── siamese_net.py # GNN to match graphs
├── toolbox
|   └── optimizer and losses selectors
|   └── logger.py  # keeping track of most results during training
|   └── metrics.py # computing scores
|   └── losses.py  # computing losses
|   └── optimizer.py # optimizers
|   └── utility.py
|   └── maskedtensor.py # Tensor-like class to handle batches of graphs of different sizes
├── commander.py # main file from the project serving for calling all necessary functions for training and testing
├── trainer.py # pipelines for training and validation
├── eval.py # testing models

Dependencies

Dependencies are listed in requirements.txt. To install, run

pip install -r requirements.txt

Training

Run the main file commander.py with the command train

python train commander.py

To change options, use Sacred command-line interface and see default.yaml for the configuration structure. For instance,

python commander.py train with cpu=No data.generative_model=Regular train.epoch=10 

You can also copy default.yaml and modify the configuration parameters there. Loading the configuration in other.yaml (or other.json) can be done with

python commander.py train with other.yaml

See Sacred documentation for an exhaustive reference.

To save logs to Neptune, you need to provide your own API key via the dedicated environment variable.

The model is regularly saved in the folder runs.

Evaluating

There are two ways of evaluating the models. If you juste ran the training with a configuration conf.yaml, you can simply do,

python commander.py eval with conf.yaml

You can omit with conf.yaml if you are using the default configuartion.

If you downloaded a model with a config file from here, you can edit the section test_data of this config if you wish and then run,

python commander.py eval with /path/to/config model_path=/path/to/model.pth.tar
Comments
  • need 2 different model_path variables

    need 2 different model_path variables

    When not starting anew, model_path_load should contain the path to the model. model_path should then be the path to save the new learned model. Right now, the model_path_load is used for the evaluation!

    bug 
    opened by mlelarge 1
  • Unify the different problems

    Unify the different problems

    The previous commands for training and evaluation should still work the same, even though it uses a different configuration file at the beginning. It still uses Sacred the same way for the commander.py. The problem can be switched from the config file. The main change in the previous code is the use of the Helper class (in toolbox/helper.py) which is the class that coordinates each problem (see the file for further info).

    Also added the 'article_commander.py' which generates the data for comparison between planted problems and the corresponding NP-problem, which works in a similar way to commander.py.

    opened by MauTrib 1
  • Reorganize a commander.py

    Reorganize a commander.py

    This PR tries to unify the train and eval CLI wth the use of a single config file. See default.yaml for the result and the README for more information. Please, don't hesitate to comment on changes you don't approve or dislike. Bests,

    opened by wazizian 1
  • Fix convention for tensor shapes

    Fix convention for tensor shapes

    Hi! I am modifying the code so the same convention for the shape of tensors is used everywhere (see PR). But I'm not sure we have chosen the right one. You and we have chosen(bs, n_vertices, n_vertices, features)but Marron preferred (bs, features, n_vertices, n_vertices). Indeed the later matches Pytorch's convention for images and so makes convolutions natural. In one case, MLPs and blocks have to be adapted and in the other, it is the data generation process, so it is a similar amount of work. What do you think ? Bests, Waïss

    opened by wazizian 1
  • added normalization + embeddings

    added normalization + embeddings

    I modified slightly the architecture and training seems much better and more stable. I added a first embedding layer and BN after multiplication and at the output.

    opened by mlelarge 0
Releases(QAP)
  • QAP(Jan 21, 2021)

    Config: "data": {"num_examples_train": 20000, "num_examples_val": 1000, "generative_model": "Regular", "noise_model": "ErdosRenyi", "edge_density": 0.2, "n_vertices": 50, "vertex_proba": 1.0, "noise": 0.15, "path_dataset": "dataset"}, "train": {"epoch": 50, "batch_size": 32, "lr": 0.0001, "scheduler_step": 5, "scheduler_decay": 0.9, "print_freq": 100, "loss_reduction": "mean"}, "arch": {"arch": "Siamese_Model", "model_name": "Simple_Node_Embedding", "num_blocks": 2, "original_features_num": 2, "in_features": 64, "out_features": 64, "depth_of_mlp": 3}

    Source code(tar.gz)
    Source code(zip)
    config.json(601 bytes)
    model_best.pth.tar(333.26 KB)
[NeurIPS 2021] ORL: Unsupervised Object-Level Representation Learning from Scene Images

Unsupervised Object-Level Representation Learning from Scene Images This repository contains the official PyTorch implementation of the ORL algorithm

Jiahao Xie 55 Dec 03, 2022
AttentionGAN for Unpaired Image-to-Image Translation & Multi-Domain Image-to-Image Translation

AttentionGAN-v2 for Unpaired Image-to-Image Translation AttentionGAN-v2 Framework The proposed generator learns both foreground and background attenti

Hao Tang 530 Dec 27, 2022
Implementation of "Unsupervised Domain Adaptive 3D Detection with Multi-Level Consistency"

Unsupervised Domain Adaptive 3D Detection with Multi-Level Consistency (ICCV2021) Paper Link: https://arxiv.org/abs/2107.11355 This implementation bui

32 Nov 17, 2022
A simple code to convert image format and channel as well as resizing and renaming multiple images.

Rename-Resize-and-convert-multiple-images A simple code to convert image format and channel as well as resizing and renaming multiple images. This cod

Happy N. Monday 3 Feb 15, 2022
The code for the CVPR 2021 paper Neural Deformation Graphs, a novel approach for globally-consistent deformation tracking and 3D reconstruction of non-rigid objects.

Neural Deformation Graphs Project Page | Paper | Video Neural Deformation Graphs for Globally-consistent Non-rigid Reconstruction Aljaž Božič, Pablo P

Aljaz Bozic 134 Dec 16, 2022
통일된 DataScience 폴더 구조 제공 및 가상환경 작업의 부담감 해소

Lucas coded by linux shell 목차 Mac버전 CookieCutter (autoenv) 1.How to Install autoenv 2.폴더 진입 시, activate 구현하기 3.폴더 탈출 시, deactivate 구현하기 4.Alias 설정하기 5

ello 3 Feb 21, 2022
Adversarial Learning for Semi-supervised Semantic Segmentation, BMVC 2018

Adversarial Learning for Semi-supervised Semantic Segmentation This repo is the pytorch implementation of the following paper: Adversarial Learning fo

Wayne Hung 464 Dec 19, 2022
Torch Implementation of "Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network"

Photo-Realistic-Super-Resoluton Torch Implementation of "Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network" [Paper]

Harry Yang 199 Dec 01, 2022
PyTorch implementation code for the paper MixCo: Mix-up Contrastive Learning for Visual Representation

How to Reproduce our Results This repository contains PyTorch implementation code for the paper MixCo: Mix-up Contrastive Learning for Visual Represen

opcrisis 46 Dec 15, 2022
Tracking Progress in Question Answering over Knowledge Graphs

Tracking Progress in Question Answering over Knowledge Graphs Table of contents Question Answering Systems with Descriptions The QA Systems Table cont

Knowledge Graph Question Answering 47 Jan 02, 2023
Multi-objective constrained optimization for energy applications via tree ensembles

Multi-objective constrained optimization for energy applications via tree ensembles

C⚙G - Imperial College London 1 Nov 19, 2021
Deep Learning (with PyTorch)

Deep Learning (with PyTorch) This notebook repository now has a companion website, where all the course material can be found in video and textual for

Alfredo Canziani 6.2k Jan 07, 2023
improvement of CLIP features over the traditional resnet features on the visual question answering, image captioning, navigation and visual entailment tasks.

CLIP-ViL In our paper "How Much Can CLIP Benefit Vision-and-Language Tasks?", we show the improvement of CLIP features over the traditional resnet fea

310 Dec 28, 2022
An official implementation of the paper Exploring Sequence Feature Alignment for Domain Adaptive Detection Transformers

Sequence Feature Alignment (SFA) By Wen Wang, Yang Cao, Jing Zhang, Fengxiang He, Zheng-jun Zha, Yonggang Wen, and Dacheng Tao This repository is an o

WangWen 79 Dec 24, 2022
Yoloxkeypointsegment - An anchor-free version of YOLO, with a simpler design but better performance

Introduction 关键点版本:已完成 全景分割版本:已完成 实例分割版本:已完成 YOLOX is an anchor-free version of

23 Oct 20, 2022
Pytorch implementation of DeePSiM

Pytorch implementation of DeePSiM

1 Nov 05, 2021
Tensorflow implementation of DeepLabv2

TF-deeplab This is a Tensorflow implementation of DeepLab, compatible with Tensorflow 1.2.1. Currently it supports both training and testing the ResNe

Chenxi Liu 21 Sep 27, 2022
learning and feeling SLAM together with hands-on-experiments

modern-slam-tutorial-python Learning and feeling SLAM together with hands-on-experiments 😀 😃 😆 Dependencies Most of the examples are based on GTSAM

Giseop Kim 59 Dec 22, 2022
Adversarial vulnerability of powerful near out-of-distribution detection

Adversarial vulnerability of powerful near out-of-distribution detection by Stanislav Fort In this repository we're collecting replications for the ke

Stanislav Fort 9 Aug 30, 2022
Image based Human Fall Detection

Here I integrated the YOLOv5 object detection algorithm with my own created dataset which consists of human activity images to achieve low cost, high accuracy, and real-time computing requirements

UTTEJ KUMAR 12 Dec 11, 2022