# Analysis of Automatic Annotation Suggestions for Hard Discourse-Level Tasks in Expert Domains

## TL;DR

This paper explores automatic annotation suggestions for complex discourse tasks in expert domains, demonstrating improvements in annotation efficiency and quality without bias, and proposes effective methods for continuous model enhancement.

## Contribution

It introduces a new dataset for epistemic activity annotation, analyzes suggestion effects, and compares methods for ongoing model improvement in expert domain tasks.

## Key findings

- Suggestions improve annotation speed and accuracy
- No significant biases introduced by suggestions
- Effective continuous model adjustment methods identified

## Abstract

Many complex discourse-level tasks can aid domain experts in their work but require costly expert annotations for data creation. To speed up and ease annotations, we investigate the viability of automatically generated annotation suggestions for such tasks. As an example, we choose a task that is particularly hard for both humans and machines: the segmentation and classification of epistemic activities in diagnostic reasoning texts. We create and publish a new dataset covering two domains and carefully analyse the suggested annotations. We find that suggestions have positive effects on annotation speed and performance, while not introducing noteworthy biases. Envisioning suggestion models that improve with newly annotated texts, we contrast methods for continuous model adjustment and suggest the most effective setup for suggestions in future expert tasks.