Table2Vec: Neural Word and Entity Embeddings for Table Population and Retrieval.

This repository contains resources developed within the following paper:

L. Deng, S. Zhang, and K. Balog. Table2Vec: Neural Word and Entity Embeddings for Table Population and Retrieval. In: Proceedings of SIGIR 2019, July 2019.

Run files

You can download a complete version of the runfiles here.

Test collection

The table corpus is WikiTables, which comprises 1.6M tables extracted from Wikipedia. We proproceeed it and make it public downloadable here.

The query file and code of table population can be found in SIGIR2017

The query and feature files of table search can be found in WWW2018

The evaluation scores are reported using trec_eval.

Note

Due to the size, the run files for row and column population are only accessiable per request.

Citation

@inproceedings{Zhang:2018:AHT,
    author = {Deng, Li, Zhang, Shuo and Balog, Krisztian},
    title = {Table2Vec: Neural Word and Entity Embeddings for Table Population and Retrieval},
    booktitle = {Proceedings of SIGIR 2019},
    year = {2019},
}

Contact

If you have any questions, please contact Shuo Zhang at [email protected].