Use of LLMs in preparing accessible scientific papers

TL;DR

This paper explores the feasibility of using large language models for zero-shot conversion of arXiv papers to improve accessibility, highlighting challenges and discussing alternative solutions.

Contribution

It presents an initial attempt at zero-shot conversion of scientific papers' formatting using LLMs, revealing limitations and proposing future directions.

Findings

Conversion is possible but unreliable

Zero-shot approach has significant challenges

Alternative methods are needed for better accessibility

Abstract

Making scientific papers accessible may require reprocessing old papers to create output compliant with accessibility standards. An important step there is to convert the visual formatting to the logical one. In this report we describe our attempt at zero shot conversion of arXiv papers. Our results are mixed: while it is possible to do conversion, the reliability is not too good. We discuss alternative approaches to this problem.