Awesome

<h1 align="center"> CompGCN </h1> <h4 align="center">Composition-Based Multi-Relational Graph Convolutional Networks</h4> <p align="center"> <a href="https://iclr.cc/"><img src="http://img.shields.io/badge/ICLR-2020-4b44ce.svg"></a> <a href="https://arxiv.org/abs/1911.03082"><img src="http://img.shields.io/badge/Paper-PDF-red.svg"></a> <a href="https://iclr.cc/virtual/poster_BylA_C4tPr.html"><img src="http://img.shields.io/badge/Video-ICLR-green.svg"></a> <a href="https://medium.com/@mgalkin/knowledge-graphs-iclr-2020-f555c8ef10e3"><img src="http://img.shields.io/badge/Blog-Medium-B31B1B.svg"></a> <a href="https://github.com/malllabiisc/CompGCN/blob/master/LICENSE"> <img src="https://img.shields.io/badge/License-Apache%202.0-blue.svg"> </a> </p> <h2 align="center"> Overview of CompGCN <img align="center" src="./overview.png" alt="..."> </h2> Given node and relation embeddings, CompGCN performs a composition operation φ(·) over each edge in the neighborhood of a central node (e.g. Christopher Nolan above). The composed embeddings are then convolved with specific filters WO and WI for original and inverse relations respectively. We omit self-loop in the diagram for clarity. The message from all the neighbors are then aggregated to get an updated embedding of the central node. Also, the relation embeddings are transformed using a separate weight matrix. Please refer to the paper for details.

Dependencies

Compatible with PyTorch 1.0 and Python 3.x.
Dependencies can be installed using requirements.txt.

Dataset:

We use FB15k-237 and WN18RR dataset for knowledge graph link prediction.
FB15k-237 and WN18RR are included in the data directory.

Training model:

Install all the requirements from requirements.txt.
Execute ./setup.sh for extracting the dataset and setting up the folder hierarchy for experiments.

Commands for reproducing the reported results on link prediction:

##### with TransE Score Function
# CompGCN (Composition: Subtraction)
python run.py -score_func transe -opn sub -gamma 9 -hid_drop 0.1 -init_dim 200

# CompGCN (Composition: Multiplication)
python run.py -score_func transe -opn mult -gamma 9 -hid_drop 0.2 -init_dim 200

# CompGCN (Composition: Circular Correlation)
python run.py -score_func transe -opn corr -gamma 40 -hid_drop 0.1 -init_dim 200

##### with DistMult Score Function
# CompGCN (Composition: Subtraction)
python run.py -score_func distmult -opn sub -gcn_dim 150 -gcn_layer 2 

# CompGCN (Composition: Multiplication)
python run.py -score_func distmult -opn mult -gcn_dim 150 -gcn_layer 2 

# CompGCN (Composition: Circular Correlation)
python run.py -score_func distmult -opn corr -gcn_dim 150 -gcn_layer 2 

##### with ConvE Score Function
# CompGCN (Composition: Subtraction)
python run.py -score_func conve -opn sub -ker_sz 5

# CompGCN (Composition: Multiplication)
python run.py -score_func conve -opn mult

# CompGCN (Composition: Circular Correlation)
python run.py -score_func conve -opn corr

##### Overall BEST:
python run.py -name best_model -score_func conve -opn corr

-score_func denotes the link prediction score score function
-opn is the composition operation used in CompGCN. It can take the following values:
- sub for subtraction operation: Φ(e_s, e_r) = e_s - e_r
- mult for multiplication operation: Φ(e_s, e_r) = e_s * e_r
- corr for circular-correlation: Φ(e_s, e_r) = e_s ★ e_r
-name is some name given for the run (used for storing model parameters)
-model is name of the model `compgcn'.
-gpu for specifying the GPU to use
Rest of the arguments can be listed using python run.py -h

Citation:

Please cite the following paper if you use this code in your work.

@inproceedings{
    vashishth2020compositionbased,
    title={Composition-based Multi-Relational Graph Convolutional Networks},
    author={Shikhar Vashishth and Soumya Sanyal and Vikram Nitin and Partha Talukdar},
    booktitle={International Conference on Learning Representations},
    year={2020},
    url={https://openreview.net/forum?id=BylA_C4tPr}
}

For any clarification, comments, or suggestions please create an issue or contact Shikhar.