CodeDistiller: Automatically Generating Code Libraries for Scientific Coding Agents

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Jansen, Peter, Hassan, Samiah, Narasimha, Pragnya
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913131338399744
author Jansen, Peter
Hassan, Samiah
Narasimha, Pragnya
author_facet Jansen, Peter
Hassan, Samiah
Narasimha, Pragnya
contents Automated Scientific Discovery (ASD) systems can help automatically generate and run code-based experiments, but their capabilities are limited by the code they can reliably generate from parametric knowledge alone. As a result, current systems either mutate a small number of manually-crafted experiment examples, or operate solely from parametric knowledge, limiting quality and reach. We introduce CodeDistiller, a system that automatically distills large collections of scientific Github repositories into a vetted library of working domain-specific code examples, allowing ASD agents to expand their capabilities without manual effort. Using a combination of automatic and domain-expert evaluation on 250 materials science repositories, we find the best model is capable of producing functional examples for 74% of repositories, while our downstream evaluation shows an ASD agent augmented with a CodeDistiller generated library produces more accurate, complete, and scientifically sound experiments than an agent with only general materials-science code examples. We also evaluate LLM-as-a-judge ratings against domain-expert ratings in an A/B testing paradigm, finding moderate agreement and suggesting that inexpensive proxy metrics may be feasible for evaluating scientific discovery systems at scale.
format Preprint
id arxiv_https___arxiv_org_abs_2512_01089
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CodeDistiller: Automatically Generating Code Libraries for Scientific Coding Agents
Jansen, Peter
Hassan, Samiah
Narasimha, Pragnya
Artificial Intelligence
Automated Scientific Discovery (ASD) systems can help automatically generate and run code-based experiments, but their capabilities are limited by the code they can reliably generate from parametric knowledge alone. As a result, current systems either mutate a small number of manually-crafted experiment examples, or operate solely from parametric knowledge, limiting quality and reach. We introduce CodeDistiller, a system that automatically distills large collections of scientific Github repositories into a vetted library of working domain-specific code examples, allowing ASD agents to expand their capabilities without manual effort. Using a combination of automatic and domain-expert evaluation on 250 materials science repositories, we find the best model is capable of producing functional examples for 74% of repositories, while our downstream evaluation shows an ASD agent augmented with a CodeDistiller generated library produces more accurate, complete, and scientifically sound experiments than an agent with only general materials-science code examples. We also evaluate LLM-as-a-judge ratings against domain-expert ratings in an A/B testing paradigm, finding moderate agreement and suggesting that inexpensive proxy metrics may be feasible for evaluating scientific discovery systems at scale.
title CodeDistiller: Automatically Generating Code Libraries for Scientific Coding Agents
topic Artificial Intelligence
url https://arxiv.org/abs/2512.01089