This is the official code base of the paper

Neural-Symbolic Models for Logical Queries on Knowledge Graphs

Zhaocheng Zhu, Mikhail Galkin, Zuobai Zhang, Jian Tang

GNN-QE is a neural-symbolic model for answering multi-hop logical queries on knowledge graphs. Given a multi-hop logical query, GNN-QE first decomposes it into 4 basic operations over fuzzy sets, and then executes the operations with graph neural networks and fuzzy logic operations.

Additionally, the intermediate variables in GNN-QE are interpretable, and one can visualize them to better understand the multi-hop reasoning process.

The dependencies can be installed via either conda or pip. GNN-QE is compatible with Python 3.7/3.8/3.9 and PyTorch >= 1.8.0.

conda install torchdrug pytorch cudatoolkit -c milagraph -c pytorch -c pyg
conda install easydict pyyaml -c conda-forge

pip install torchdrug torch
pip install easydict pyyaml

To run GNN-QE, use the following command. Alternatively, you may specify --gpus null to run GNN-QE on a CPU. All the datasets will be automatically downloaded in the code.

python script/run.py -c config/fb15k237.yaml --gpus [0]

We provide the hyperparameters for each experiment in a separate configuration file. The hyperparameters are designed for 32GB GPUs, but you may adjust the batch size to fit a smaller GPU memory. All the configuration files can be found in config/*.yaml.

To run GNN-QE with multiple GPUs or multiple machines, use the following commands

python -m torch.distributed.launch --nproc_per_node=4 script/run.py -c config/fb15k237.yaml --gpus [0,1,2,3]

python -m torch.distributed.launch --nnodes=4 --nproc_per_node=4 script/run.py -c config/fb15k237.yaml --gpus [0,1,2,3,0,1,2,3,0,1,2,3,0,1,2,3]

Once you have models trained on FB15k-237, you can visualize the intermediate variables with the following line. The visualizations will be output in a new experiment directory as png files. Please replace the checkpoint with your own path.

python script/visualize.py -c config/fb15k237_visualize.yaml --checkpoint /path/to/gnn-qe/experiment/model_epoch_10.pth

Generally, GNN-QE is a GNN-agnostic framework for answering logical queries. You may plug any GNN model into GNN-QE. To do so, just implement the convolution layer in gnnqe/layer.py and the GNN model in gnnqe/gnn.py. The GNN model is expected to have the following signature

def forward(self, graph, input, all_loss=None, metric=None):
    ...
    return {
        "node_feature": node_feature,
    }

where the arguments and the return value are

• graph (data.Graph): the knowledge graph with graph.query being the query embeddings of shape (batch_size, input_dim).
• input (Tensor): input tensor of shape (|V|, batch_size, input_dim)
• all_loss (Tensor): a scalar tensor that accumulates losses during training
• metric (Tensor): a dict that stores any scalar information for logging during training
• node_feature (Tensor): node feature of shape (|V|, batch_size, output_dim)

For how to implement GNN models in TorchDrug, please refer to these tutorials

• Graph Neural Network Layers
• Customize Models & Tasks

1. The code is stuck at the beginning of epoch 0.

   This is probably because the JIT cache is broken. Try rm -r ~/.cache/torch_extensions/* and run the code again.

If you find this project useful in your research, please cite the following paper

@inproceedings{zhu2022neural,
  title={Neural-Symbolic Models for Logical Queries on Knowledge Graphs},
  author={Zhu, Zhaocheng and Galkin, Mikhail and Zhang, Zuobai and Tang, Jian},
  booktitle={Proceedings of the 39th International Conference on Machine Learning},
  pages={27454--27478},
  volume={162},
  year={2022},
}