Zero-shot User Intent Detection via Capsule Neural Networks (PyTorch Implementation)

This repository implements a capsule model named IntentCapsNet-ZSL on the SNIPS-NLU dataset with PyTorch (extension of Tensorflow version)

The official Tensorflow version is available: https://github.com/congyingxia/ZeroShotCapsule

Please see the following paper for the details:

Congying Xia*, Chenwei Zhang*, Xiaohui Yan, Yi Chang, Philip S. Yu. Zero-shot User Intent Detection via Capsule Neural Networks. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2018. (* equally contributed)

https://arxiv.org/abs/1809.00385

Requirements

Python 2.7.12

torch 1.0.1

Numpy

Gensim

Sklearn

Usage

python main.py

If you find the code useful, please cite the paper.

@article{xia2018zero,
  title={Zero-shot User Intent Detection via Capsule Neural Networks},
  author={Xia, Congying and Zhang, Chenwei and Yan, Xiaohui and Chang, Yi and Yu, Philip S},
  journal={arXiv preprint arXiv:1809.00385},  
  year={2018}
}

Acknowledgements

https://github.com/congyingxia/ZeroShotCapsule

https://github.com/soskek/dynamic_routing_between_capsules

https://github.com/ExplorerFreda/Structured-Self-Attentive-Sentence-Embedding