From News to Medical: Cross-domain Discourse Segmentation
TL
TL;DR
This work annotates the first high-quality small-scale medical corpus in English with discourse segments and analyzes how well news-trained segmenters perform on this domain.
Abstract
The first step in discourse analysis involves dividing a text into segments. We annotate the first high-quality small-scale medical corpus in English with discourse segments and analyze how well news-trained segmenters perform on this domain. While we expectedly find a drop in performance, the nature of the segmentation errors suggests some problems can be addressed earlier in the pipeline, while others would require expanding the corpus to a trainable size to learn the nuances of the medical domain.
References30 items
1
What do you learn from context? Probing for sentence structure in contextualized word representations
2
Toward Fast and Accurate Neural Discourse Segmentation
3
SciDTB: Discourse Dependency TreeBank for Scientific Abstracts
4
Deep Contextualized Word Representations
5