LLMs for industrial databases: an agro-food production plant use case

Lacramioara Dranca, Pablo Donate, Julio A. Sanguesa, Piedad Garrido, Vicente Torres-Sanz, Francisco J. Martinez · 2026

Read LLMs for industrial databases: an agro-food production plant use case by Lacramioara Dranca, Pablo Donate, Julio A. Sanguesa, Piedad Garrido, Vicente Torres-Sanz, Francisco J. Martinez (2026) 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/pmc-13260371.txt.

Description

Lacramioara Dranca and co-authors built in 2026 a system that lets staff question an industrial database in plain Spanish instead of code. The test site was a working agro-food plant in Spain that generates its own energy. Its time-series database uses special query languages that plant experts rarely know.

A knowledge graph maps the database layout onto the plant's real equipment, giving the system shared ground truth. A small fine-tuned language model turns questions into queries, which a checker then validates for safety and sense. Training data came from a large teacher model filtered through automatic checks, avoiding hand labeling. On 56 test questions, small models wrote valid queries reliably, yet fewer ran cleanly and fewer still returned correct answers. The authors locate the remaining gap in meaning, not grammar.

Catalog record

CREATORLacramioara Dranca, Pablo Donate, Julio A. Sanguesa, Piedad Garrido, Vicente Torres-Sanz, Francisco J. Martinez
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE163 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13260371

Browse the full library · About the Tower of Babel