Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/xovee/ccgl

TKDE 22. CCCL: Contrastive Cascade Graph Learning.
https://github.com/xovee/ccgl

acm aps data-augmentation dblp graph information-cascade information-diffusion popularity-prediction self-supervised-learning twitter weibo

Last synced: about 12 hours ago
JSON representation

TKDE 22. CCCL: Contrastive Cascade Graph Learning.

Host: GitHub
URL: https://github.com/xovee/ccgl
Owner: Xovee
License: mit
Created: 2020-08-31T12:47:29.000Z (about 4 years ago)
Default Branch: master
Last Pushed: 2024-03-12T09:57:23.000Z (8 months ago)
Last Synced: 2024-03-12T11:04:53.564Z (8 months ago)
Topics: acm, aps, data-augmentation, dblp, graph, information-cascade, information-diffusion, popularity-prediction, self-supervised-learning, twitter, weibo
Language: Python
Homepage: https://www.xoveexu.com
Size: 13.5 MB
Stars: 28
Watchers: 3
Forks: 5
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

README

        # CCGL: Contrastive Cascade Graph Learning

![](https://img.shields.io/badge/TKDE-22-blue)

![](https://img.shields.io/badge/python-3.7-green)

![](https://img.shields.io/badge/tensorflow-2.3-green)

![](https://img.shields.io/badge/cudatoolkit-10.1-green)

![](https://img.shields.io/badge/cudnn-7.6.5-green)

This repo provides a reference implementation of Contrastive Cascade Graph Learning (**CCGL**) framework as described in the paper:

> CCGL: Contrastive Cascade Graph Learning   

> [Xovee Xu](https://www.xoveexu.com), [Fan Zhou](https://dblp.org/pid/63/3122-2.html), [Kunpeng Zhang](http://www.terpconnect.umd.edu/~kpzhang/), and [Siyuan Liu](https://directory.smeal.psu.edu/sxl68)  

> IEEE Transactions on Knowledge and Data Engineering (TKDE), vol. 35, no. 5, pp. 4539--4554, May 2023   

> [arXiv:2107.12576](https://arxiv.org/abs/2107.12576)  

> https://doi.org/10.1109/TKDE.2022.3151829  

## Dataset

You can download all five datasets (Weibo, Twitter, ACM, APS, and DBLP) via any one of the following links:

Google Drive|Dropbox|Onedrive|Baidu Netdisk

:---:|:---:|:---:|:---:

||| `trqg`

## Environmental Settings

Our experiments are conducted on Ubuntu 20.04, a single NVIDIA 1080Ti GPU, 48GB RAM, and Intel i7 8700K. CCGL is implemented by `Python 3.7`, `TensorFlow 2.3`, `Cuda 10.1`, and `Cudnn 7.6.5`.

Create a virtual environment and install GPU-support packages via [Anaconda](https://www.anaconda.com/):

```shell

# create virtual environment

conda create --name=ccgl python=3.7 cudatoolkit=10.1 cudnn=7.6.5

# activate virtual environment

conda activate ccgl

# install other dependencies

pip install -r requirements.txt

```

## Usage

Here we take Weibo dataset as an example to demonstrate the usage.

### Preprocess

Step 1: divide, filter, generate labeled and unlabeled cascades:

```shell

cd ccgl

# labeled cascades

python src/gene_cas.py --input=./datasets/weibo/ --unlabel=False

# unlabeled cascades

python src/gene_cas.py --input=./datasets/weibo/ --unlabel=True

```

Step 2: augment both labeled and unlabeled cascades (here we use the `AugSIM` strategy):

```shell

python src/augmentor.py --input=./datasets/weibo/ --aug_strategy=AugSIM

```

Step 3: generate cascade embeddings:

```shell

python src/gene_emb.py --input=./datasets/weibo/ 

```

### Pre-training

```shell

python src/pre_training.py --name=weibo-0 --input=./datasets/weibo/ --projection_head=4-1

```

The saved pre-training model is named as `weibo-0`. 

### Fine-tuning

```shell

python src/fine_tuning.py --name=weibo-0 --num=0 --input=./datasets/weibo/ --projection_head=4-1

```

Here we load the pre-trained model `weibo-0` and save the teacher network as `weibo-0-0`.

### Distillation

```shell

python src/distilling.py --name=weibo-0-0 --num=0 --input=./datasets/weibo/ --projection_head=4-1

```

Here we load the teacher network `weibo-0-0` and save the student network as `weibo-0-0-student-0`. 

### (Optional) Run the Base model

```shell

python src/base_model.py --input=./datasets/weibo/ 

```

## CCGL model weights

We provide pre-trained, fine-tuned, and distilled CCGL model weights. Please see details in the following table. 

Model|Dataset|Label Fraction|Projection Head|MSLE|Weights

:---|:---|:---|:---|:---|:---

Pre-trained CCGL model|Weibo|100%|4-1|-|[Download](./results/pre_training_weight/weibo-100.h5)

Pre-trained CCGL model|Weibo|10%|4-4|-|[Download](./results/pre_training_weight/weibo-10.h5)

Pre-trained CCGL model|Weibo|1%|4-3|-|[Download](./results/pre_training_weight/weibo-1.h5)

Fine-tuned CCGL model|Weibo|100%|4-1|2.70|[Download](./results/fine_tuning_weight/weibo-100-0.h5)

Fine-tuned CCGL model|Weibo|10%|4-4|2.87|[Download](./results/fine_tuning_weight/weibo-10-0.h5)

Fine-tuned CCGL model|Weibo|1%|4-3|3.30|[Download](./results/fine_tuning_weight/weibo-1-0.h5)

Load weights into the model:

```python

# construct model, carefully check projection head designs:

# use different number of Dense layers

...

# load weights for fine-tuning, distillation, or evaluation

model.load_weights(weight_path)

```

Check `src/fine_tuning.py` and `src/distilling.py` for *weights loading* examples.

## Default hyper-parameter settings

Unless otherwise specified, we use following default hyper-parameter settings.

Param|Value|Param|Value

:---|---:|:---|---:

Augmentation strength|0.1|Pre-training epochs|30

Augmentation strategy|AugSIM|Projection Head (100%)|4-1

Batch size|64|Projection Head (10%)|4-4

Early stopping patience|20|Projection Head (1%)|4-3

Embedding dimension|64|Model size|128 (4x)

Learning rate|5e-4|Temperature|0.1

## Change Logs

- Jul 21, 2021: fix a bug and some annotations

## Cite

If you find our paper & code are useful for your research, please consider citing us 😘:

```bibtex

@article{xu2022ccgl, 

  author = {Xovee Xu and Fan Zhou and Kunpeng Zhang and Siyuan Liu}, 

  title = {{CCGL}: Contrastive Cascade Graph Learning}, 

  journal = {IEEE Transactions on Knowledge and Data Engineering (TKDE)},

  volume = {35},

  number = {5},

  pages = {4539--4554},

  year = {2022},

  doi = {10.1109/TKDE.2022.3151829}, 

}

```

We also have a [survey paper](https://dl.acm.org/doi/10.1145/3433000?cid=99659687041) you might be interested:

```bibtex

@article{zhou2021survey,

  author = {Fan Zhou and Xovee Xu and Goce Trajcevski and Kunpeng Zhang}, 

  title = {A Survey of Information Cascade Analysis: Models, Predictions, and Recent Advances}, 

  journal = {ACM Computing Surveys (CSUR)}, 

  volume = {54},

  number = {2},

  year = {2021},

  articleno = {27},

  numpages = {36},

  doi = {10.1145/3433000},

}

```

## Acknowledgment

We would like to thank [Xiuxiu Qi](https://qhemu.github.io/xiuxiuqi/), [Ce Li](https://github.com/celi52), [Qing Yang](https://www.linkedin.com/in/庆-杨-43ba1a142), and Wenxiong Li for sharing their computing resources and help us to test the codes. We would also like to show our gratitude to the authors of [SimCLR](https://github.com/google-research/simclr) (and [Sayak Paul](https://github.com/sayakpaul)), [node2vec](https://github.com/eliorc/node2vec), [DeepHawkes](https://github.com/CaoQi92/DeepHawkes), and others, for sharing their codes and datasets. 

## Contact

For any questions please open an issue or drop an email to: `xovee at live.com`