Skip to content

Code and Task setup of the paper "One Representation per Word - Does it make Sense for Composition?"

Notifications You must be signed in to change notification settings

tttthomasssss/sense2017

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

sense2017

Task setup of the paper "One Representation per Word - Does it make Sense for Composition?"

This package re-creates the dataset as used in the paper.

If you use this dataset in your research please use the following citation entry:

@InProceedings{kober-EtAl:2017:SENSE2017,
  author    = {Kober, Thomas  and  Weeds, Julie  and  Wilkie, John  and  Reffin, Jeremy  and  Weir, David},
  title     = {One Representation per Word - Does it make Sense for Composition?},
  booktitle = {Proceedings of the 1st Workshop on Sense, Concept and Entity Representations and their Applications},
  month     = {April},
  year      = {2017},
  address   = {Valencia, Spain},
  publisher = {Association for Computational Linguistics},
  pages     = {79--90},
  abstract  = {In this paper, we investigate whether an a priori disambiguation of word senses
	is strictly necessary or whether the meaning of a word in context can be
	disambiguated through composition alone. We evaluate the performance of
	off-the-shelf single-vector and multi-sense vector models on a benchmark phrase
	similarity task and a novel task for word-sense discrimination. We find that
	single-sense vector models perform as well or better than multi-sense vector
	models despite arguably less clean elementary representations. Our findings
	furthermore show that simple composition functions such as pointwise addition
	are able to recover sense specific information from a single-sense vector model
	remarkably well.},
  url       = {http://www.aclweb.org/anthology/W17-1910}
}

Please email the first author if you have any queries with regards to this task.

Prerequisits

Installation

If the repository is cloned, the code should run without the need for running the setup files, if you want to install it anyway do the following:

cd /path/to/sense2017
python setup.py install

Requirements

  • beautifulsoup4>=4.5.1
  • fuzzywuzzy>=0.15.0
  • joblib>=0.9.2
  • nltk>=3.0.2
  • requests>=2.11.1

To install all requirements without installing sense2017 you can run:

cd /path/to/sense2017
pip install -r requirements.txt

Usage

To re-create our task run the following:

python -m sense2017.create_dataset -oxid <OXFORD-APP-ID> -oxkey <OXFORD-APP-KEY> -cokey <COLLINS-API-KEY> -op /path/to/output/folder

About

Code and Task setup of the paper "One Representation per Word - Does it make Sense for Composition?"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages