Incorporating Both Distributional and Relational Semantics in Word Representations
Abstract
We investigate the hypothesis that word representations ought to incorporate both distributional and relational semantics. To this end, we employ the Alternating Direction Method of Multipliers (ADMM), which flexibly optimizes a distributional objective on raw text and a relational objective on WordNet. Preliminary results on knowledge base completion, analogy tests, and parsing show that word representations trained on both objectives can give improvements in some cases.
Cite
Text
Fried and Duh. "Incorporating Both Distributional and Relational Semantics in Word Representations." International Conference on Learning Representations, 2015.Markdown
[Fried and Duh. "Incorporating Both Distributional and Relational Semantics in Word Representations." International Conference on Learning Representations, 2015.](https://mlanthology.org/iclr/2015/fried2015iclr-incorporating/)BibTeX
@inproceedings{fried2015iclr-incorporating,
title = {{Incorporating Both Distributional and Relational Semantics in Word Representations}},
author = {Fried, Daniel and Duh, Kevin},
booktitle = {International Conference on Learning Representations},
year = {2015},
url = {https://mlanthology.org/iclr/2015/fried2015iclr-incorporating/}
}