Paper2Code: Automating Code Generation from Scientific Papers in Machine Learning

Minju Seo, Jinheon Baek, Seongyun Lee, Sung Ju Hwang · 2025

Read Paper2Code: Automating Code Generation from Scientific Papers in Machine Learning by Minju Seo, Jinheon Baek, Seongyun Lee, Sung Ju Hwang (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-2504-17192.txt.

Introduction

Minju Seo, Jinheon Baek, Seongyun Lee, and Sung Ju Hwang present PaperCoder, a team of AI agents that turns machine learning papers into working code stores. Many papers lack shared code, so rebuilding their results is slow and heavy work. Their work came out in twenty twenty-five.

The frame works in three stages. Planning maps the road ahead, draws the system design, and lists file links and settings. Analysis digs into the fine steps needed to build each part. Code writing then yields modular code that respects those links. Each stage runs through agents made for its role. Author judged tests show the code is true to the papers.

Catalog record

CREATORMinju Seo, Jinheon Baek, Seongyun Lee, Sung Ju Hwang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE159 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-17192

Browse the full library · About the Tower of Babel