WIPS : PyTorch implementation of the paper "Representation Learning with Weighted Inner Product for Universal Approximation of General Similarities"

WIPS is an open source implementation of the paper "Representation Learning with Weighted Inner Product for Universal Approximation of General Similarities." IJCAI-19.

Requirements & Test Environment

• python3
• pytorch=1.1.0
• scikit-learn
• tqdm
• nltk
• gensim
• numpy
• Tested on CentOS Linux release 7.4.1708

Acknowledgements

The implementation is based on SIPS, see also the implementation of the AISTATS-19 paper "Graph Embedding with Shifted Inner Product Similarity and Its Improved Approximation Capability". Some of the code is also based on Facebook's poincare-embeddings, see also their implementation of the NIPS-17 paper "Poincaré Embeddings for Learning Hierarchical Representations".

Reference

• BibTeX

If you find this code useful for your research, please cite the following paper in your publication:

@inproceedings{ijcai2019-699,
  title     = {Representation Learning with Weighted Inner Product for Universal Approximation of General Similarities},
  author    = {Kim, Geewook and Okuno, Akifumi and Fukui, Kazuki and Shimodaira, Hidetoshi},
  booktitle = {Proceedings of the Twenty-Eighth International Joint Conference on
               Artificial Intelligence, {IJCAI-19}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},             
  pages     = {5031--5038},
  year      = {2019},
  month     = {7},
  doi       = {10.24963/ijcai.2019/699},
  url       = {https://doi.org/10.24963/ijcai.2019/699},
}

License

This code is licensed under CC-BY-NC 4.0.