Adapting Biomedical Abstracts into Plain language using Large Language Models

Haritha Gangavarapu 2, Giridhar Kaushik Ramachandran 1, Kevin Lybarger 1, Meliha Yetisgen 3, Özlem Uzuner 1 · 2025

Read Adapting Biomedical Abstracts into Plain language using Large Language Models by Haritha Gangavarapu 2, Giridhar Kaushik Ramachandran 1, Kevin Lybarger 1, Meliha Yetisgen 3, Özlem Uzuner 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-15700.txt.

Description

Haritha Gangavarapu and co-authors describe, in a 2025 paper, systems that rewrite medical research abstracts in plain language for the public. Most people lack the health literacy to use dense medical writing well. The work was part of a shared task that pairs PubMed abstracts with health questions from a public site. The goal is a sentence-by-sentence rewrite that a lay reader can follow.

The team fine-tuned strong open language models made for dialogue, and also tested a top closed model. Systems were scored on how simple the rewrites were and on how true they stayed to the source. Their best system, built on GPT-4, ranked first on simplicity and third on accuracy among the teams taking part. The results show that plain rewrites can stay faithful, a key need for public health reading.

Catalog record

CREATORHaritha Gangavarapu 2, Giridhar Kaushik Ramachandran 1, Kevin Lybarger 1, Meliha Yetisgen 3, Özlem Uzuner 1
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-15700

Browse the full library · About the Tower of Babel