# Text Simplification of Scientific Texts for Non-Expert Readers

**Authors:** Björn Engelmann, Fabian Haak, Christin Katharina Kreutz, Narjes Nikzad Khasmakhi, Philipp Schaer  
**arXiv:** [2307.03569](https://arxiv.org/abs/2307.03569)  
**Date:** 2023-07-07  
**Summary page:** https://www.summarizepaper.com/en/paper/2307.03569v1/

## Key points

- Text simplification is crucial for catering to diverse reading levels
- Reading comprehension can vary based on language proficiency, cognitive abilities, and prior knowledge
- Simplifying scientific abstracts helps non-experts understand essential information
- SimpleText lab focuses on advancing text simplification for non-experts through Task 3
- Authors utilized T5 models, PEGASUS, and ChatGPT to simplify complex texts
- Research accepted at SimpleText@CLEF'23 with 12 pages, 1 figure, and 4 tables
- Significance of text simplification in enhancing access to scientific information for non-expert readers

## Summary

In their paper titled "Text Simplification of Scientific Texts for Non-Expert Readers," authors Björn Engelmann, Fabian Haak, Christin Katharina Kreutz, Narjes Nikzad Khasmakhi, and Philipp Schaer delve into the importance of text simplification in catering to the diverse reading levels of individuals. They highlight how reading comprehension can vary based on factors such as language proficiency, cognitive abilities, and prior knowledge on a subject. The authors emphasize that text simplification involves rephrasing complex texts to make them more accessible to specific target reader groups. Specifically focusing on scientific abstracts, the authors discuss how simplifying these texts can enable non-experts to grasp essential information without requiring specialized domain knowledge. is particularly crucial for audiences like cancer patients seeking information on innovative treatment options. To address this need, the SimpleText lab is dedicated to advancing the field of  for non-experts through Task 3. The authors detail their contributions in this area by presenting three runs utilizing various . Two runs are based on T5 models, one run utilizes PEGASUS, and another run employs ChatGPT with a focus on identifying complex phrases within the text. Their research was accepted at SimpleText@CLEF'23 and includes 12 pages, 1 figure, and 4 tables. Overall, this study underscores the significance of  in enhancing  to scientific information for non-expert readers. By employing advanced summarization techniques and complex phrase identification tools, the authors aim to facilitate easier comprehension of complex scientific concepts for a wider audience.