CABI has published its first artificial intelligence-ready datasets on Hugging Face, one of the world’s leading platforms for sharing machine learning models, datasets and AI resources. Developed as part of the ‘Generative AI for Agriculture (GAIA)(opens in new tab)‘ project, the datasets make CABI’s agricultural knowledge more accessible.
The Gates Foundation and UK International Development from the UK government support the GAIA project. GAIA aims to address inequalities by improving content sourcing, sharing and governance across Generative AI systems. This will help make digital advisory tools safer, more relevant and more equitable.
The first dataset, CABI Plant Health Knowledge for AI, covers Kenya, Ethiopia and India and ,according to Hugging Face "provides a collection of practical, multilingual and localised plant health advisory content from CABI, specifically processed and structured for Large Language Models (LLMs) and retrieval augmented generation (RAG) applications in agricultural advisory services." CABI has also contributed its CABI Palm Weevil Knowledge for AI to Hugging Face. That dataset combines books, journal articles, CABI Compendium, datasheets, factsheets for farmers and country-specific information. Where available, it also includes registered bioprotection products.
The team plans to make further datasets available over the coming months. A livestock dataset covering poultry and cattle is expected later this year. The Plant Health Content collection will also continue to expand, with additional plant health resources, including bioprotection product information.
Developers and researchers interested in using AI for agriculture are invited to explore the datasets by contacting CABI at enquiries@cabi.org.
Read the full press release here.