Representing text as abstract images enables image classifiers to also simultaneously classify text
Abstract
We introduce a novel method for converting text data into abstract image representations, which allows image-based processing techniques (e.g. image classification networks) to be applied to text-based comparison problems. We apply the technique to entity disambiguation of inventor names in US patents. The method involves converting text from each pairwise comparison between two inventor name records into a 2D RGB (stacked) image representation. We then train an image classification neural network to discriminate between such pairwise comparison images, and use the trained network to label each pair of records as either matched (same inventor) or non-matched (different inventors), obtaining highly accurate results. Our new text-to-image representation method could also be used more broadly for other NLP comparison problems, such as disambiguation of academic publications, or for problems that require simultaneous classification of both text and image datasets.
Cite
@article{arxiv.1908.07846,
title = {Representing text as abstract images enables image classifiers to also simultaneously classify text},
author = {Stephen M. Petrie and T'Mir D. Julius},
journal= {arXiv preprint arXiv:1908.07846},
year = {2020}
}
Comments
Minor changes in order to submit paper to a different conference (e.g. made minor changes to writing in several places and added extra data to Table 3 in order to make it clearer)