From Articles to Code: On-Demand Generation of Core Algorithms from Scientific Publications

TL;DR

This paper demonstrates that large language models can generate reliable implementations of core algorithms directly from scientific publication descriptions, potentially replacing traditional software libraries and reducing maintenance costs.

Contribution

It introduces a method to use scientific articles as specifications for LLMs to generate code, showing promising results across multiple models and algorithms.

Findings

LLMs can reproduce core algorithms with performance comparable to traditional libraries.

Scientific publications can serve as effective standalone specifications for code generation.

On-demand code generation from articles could reduce software maintenance overhead.

Abstract

Maintaining software packages imposes significant costs due to dependency management, bug fixes, and versioning. We show that rich method descriptions in scientific publications can serve as standalone specifications for modern large language models (LLMs), enabling on-demand code generation that could supplant human-maintained libraries. We benchmark state-of-the-art models (GPT-o4-mini-high, Gemini Pro 2.5, Claude Sonnet 4) by tasking them with implementing a diverse set of core algorithms drawn from original publications. Our results demonstrate that current LLMs can reliably reproduce package functionality with performance indistinguishable from conventional libraries. These findings foreshadow a paradigm shift toward flexible, on-demand code generation and away from static, human-maintained packages, which will result in reduced maintenance overhead by leveraging published articles as sufficient context for the automated implementation of analytical workflows.