Skip to yearly menu bar Skip to main content


Poster
in
Workshop: CODEML: Championing Open-source DEvelopment in Machine Learning

An LLM-Powered Tool for Enhancing Scientific Open-Source Repositories

Nikolay Nikitin · Andrey Getmanov · Zakhar Popov · Ulyanova Alekseevna · Yaroslav Aksenkin · Ilya Sokolov · Alexander Boukhanovsky

[ ] [ Project Page ]
Fri 18 Jul 2:15 p.m. PDT — 3 p.m. PDT

Abstract:

We present ISOS (Improver for Scientific Open Source), an open-source tool that uses large language models (LLMs) to improve scientific repositories. ISOS uses an agent-based architecture to automate tasks such as generating README files, docstrings and documentation, configuring CI/CD pipelines, creating tag annotations and analysing repositories to provide recommendations on best practice. Unlike code-generation tools, ISOS focuses on enhancing existing repositories. We have validated ISOS's effectiveness across diverse scientific repositories in multiple domains.

Chat is not available.