An empirical convolutional neural network approach for semantic relation classification

PyTorch implementation of Relcation Classification model described in our Neurocomputing paper [An empirical convolutional neural network approach for semantic relation classification(https://www.sciencedirect.com/science/article/pii/S0925231216000023) on the SemEval-2010 Task-8 dataset.

Steps to run the experiments

Requirements

• Python 2.7.12
• PyTorch 0.4.1
• panda 0.19.1

Datasets and word embeddings

• Dataset is already included in the directory ./SemEval2010_task8_all_data.
• Embedding file ./data/word_vecs.pkl is generated from the released word embedding file GoogleNews-vectors-negative300.bin (http://code.google.com/p/word2vec/) by Mikolov.

Training

• python train.py

Output

• The proposed anwser is outputed to the directory ./Answers.

Test

• In the directory ./SemEval2010_task8_all_data/SemEval2010_task8_scorer-v1.2
• perl semeval2010_task8_scorer-v1.2.pl ../../Answers/proposed_answer.txt TEST_KEY_19.TXT

Reference

@article{Qin2016An,
  title={An empirical convolutional neural network approach for semantic relation classification},
  author={Qin, Pengda and Xu, Weiran and Guo, Jun},
  journal={Neurocomputing},
  volume={190},
  pages={1-9},
  year={2016},
}