Skip to content
master
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
bem
 
 
 
 
 
 
 
 
 
 
 
 

README.md

Prerequisites:

  1. Tensorflow 1.10 or above
  2. Install faiss-cpu and faiss-gpu as per http://github.com/facebookresearch/faiss/blob/master/INSTALL.md.

Example Commands:

Train BEM

CUDA_VISIBLE_DEVICES=1 python -u main_wiki.py -trpr 'train' -df 1 -nb 500 -kt 'wiki_TransE.txt' -bt 'pagelink_node2vec.txt' -pos 'pair'

Prediction (Embedding Correction)

CUDA_VISIBLE_DEVICES=1 python -u main_wiki.py -trpr 'pred' -df 1 -nb 500 -kt 'wiki_TransE.txt' -bt 'pagelink_node2vec.txt' -pos 'pair'

Evaluation on the classification task

  • the raw embedding:

CUDA_VISIBLE_DEVICES=1 python -u main_eval_cl.py -df 1 -ron 'raw' -ioK 1 -dn 'wiki_net'

  • the new/corrected embedding:

CUDA_VISIBLE_DEVICES=1 python -u main_eval_cl.py -df 1 -ron 'new' -ioK 1 -dn 'wiki_net'


Issues

If you encounter any bugs or have any specific feature requests, please file an issue.


Citation

@inproceedings{ye2019bayes,
  title={Bayes EMbedding (BEM) Refining Representation by Integrating Knowledge Graphs and Behavior-specific Networks},
  author={Ye, Yuting and Wang, Xuwu and Yao, Jiangchao and Jia, Kunyang and Zhou, Jingren and Xiao, Yanghua and Yang, Hongxia},
  booktitle={Proceedings of the 28th ACM International Conference on Information and Knowledge Management},
  pages={679--688},
  year={2019}
}

About

This project targets at using Bayesian method agglomerate multiple embeddings from different sources. The paper can be found at https://arxiv.org/abs/1908.10611.

Resources

Releases

No releases published

Packages

No packages published

Languages

You can’t perform that action at this time.