Leveraging large language models to maintain a branded food product database
Pascal Hauff, Carolin Krems, Jan Kohl · 2026
Read Leveraging large language models to maintain a branded food product database by Pascal Hauff, Carolin Krems, Jan Kohl (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/PMC13273058.txt.
Description
Pascal Hauff leads a 2026 study using large language models to maintain a database of branded food products. Branded food data help track modern diets, but the data are vast, fast-changing, and uneven in quality. The team built a fully automated pipeline to collect, standardize, and enrich product data down to ingredient level.
A fine-tuned model beat human experts at parsing and mapping product data. Models without fine-tuning performed poorly, while even modest amounts of tuning data improved results a lot. The authors say language models offer a scalable way to keep food databases consistent, and can help estimate ingredient amounts and nutrients not declared on labels.
Catalog record
| CREATOR | Pascal Hauff, Carolin Krems, Jan Kohl |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50413 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |