Explainable Link Prediction for Emerging Entities in Knowledge Graphs

Acknowledgement

This repository is adopted from the source code of the paper Multi-Hop Knowledge Graph Reasoning with Reward Shaping. Lin et al., 2018 We thank the authors, especially, Xi Victoria Lin for helping us to understand their source code.

Requirements

python 3.6+
pytorch 1.4.0
tqdm 4.9.0

All experiments are run on NVIDIA Titan RTX GPUs with 24GB memory.

Data Processing

Run the following command to preprocess the datasets.

./experiment.sh configs/<dataset>.sh --process_data <gpu-ID>

<dataset> is the name of the dataset in the ./data directory. In our experiments, the three datasets used are: fb15k-237, wn18rr and nell-995. <gpu-ID> is a non-negative integer number representing the GPU index.

Model Training

For reward shaping, we used ConvE model.

1. Train the ConvE model on the training split of the data

./experiment-emb.sh configs/<dataset>-conve.sh --train <gpu-ID>

2. Train RL models (policy gradient + reward shaping)

./experiment-rs.sh configs/<dataset>-rs.sh --train <gpu-ID>

3. To train an ablated version of the model (only policy gradient)

./experiment.sh configs/<dataset>.sh --train <gpu-ID>

• Note: To train the RL models using reward shaping, make sure 1) you have pre-trained the embedding-based ConvE model and 2) set the file path pointers conve_state_dict_path to the pre-trained embedding-based models correctly in the configs/<dataset>-rs.sh or configs/<dataset>.sh files.

Evaluation

To generate the evaluation results of a trained model, simply change the --train flag in the commands above to --inference.

For example, the following command performs inference with the RL models (policy gradient + reward shaping) and prints the evaluation results (on both dev and test sets).

./experiment-rs.sh configs/<dataset>-rs.sh --inference <gpu-ID>

To print the inference paths generated by beam search during inference, use the --save_beam_search_paths flag:

./experiment-rs.sh configs/<dataset>-rs.sh --inference <gpu-ID> --save_beam_search_paths

• Note for the NELL-995 dataset:

  On this dataset we split the original training data into train.triples and dev.triples, and the final model to test has to be trained with these two files combined.

  1. To obtain the correct test set results, you need to add the --test flag to all training and inference commands.

  # You may need to adjust the number of training epochs based on the dev set development.
  ./experiment.sh configs/nell-995.sh --process_data <gpu-ID> --test
  ./experiment-emb.sh configs/nell-995-conve.sh --train <gpu-ID> --test
  ./experiment-rs.sh configs/nell-995-rs.sh --train <gpu-ID> --test
  ./experiment-rs.sh configs/nell-995-rs.sh --inference <gpu-ID> --test
  

Change the hyperparameters

To change the hyperparameters and other experiment set up, start from the configuration files.

Citation

@inproceedings{BhowmikDeMelo2020ExplainableKGReasoning,
  author    = {Bhowmik, Rajarshi and de Melo, Gerard},
  title     = {Explainable Link Prediction for Emerging Entities in Knowledge Graphs},
  booktitle = {Proceedings of ISWC 2020},
  year      = {2020},
  eventdate = {2020-11-02/06},
}