# From News to Medical: Cross-domain Discourse Segmentation

## TL;DR

This paper explores how well discourse segmentation models trained on news data perform on medical texts, highlighting challenges and potential solutions for cross-domain adaptation.

## Contribution

It introduces a new annotated medical corpus and analyzes cross-domain segmentation performance, revealing domain-specific challenges and avenues for improving segmentation models.

## Key findings

- Performance drops when applying news-trained segmenters to medical texts
- Some segmentation errors can be addressed with pipeline improvements
- Expanding the corpus is necessary to learn medical domain nuances

## 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.