AUTHOR=Omara Jonathan , Talavera Estefania , Otim Daniel , Turcza Dan , Ofumbi Emmanuel , Owomugisha Godliver 

TITLE=A field-based recommender system for crop disease detection using machine learning

JOURNAL=Frontiers in Artificial Intelligence

VOLUME=Volume 6 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2023.1010804

DOI=10.3389/frai.2023.1010804

ISSN=2624-8212

ABSTRACT=This study investigates crop disease monitoring with real time information feedback to smallholder farmers. Proper crop disease diagnosing tools and information about agricultural practices is key to growth and development in the agricultural sector. This study was piloted in a rural community of smallholder farmers having 100 farmers participating on a system that performs diagnosis on cassava diseases and provides advisory recommendation services with real-time information. In this work we deployed a crop disease diagnostic tool building on the study by the  Owomugisha and Mwebaze (2016), Makerere AI Lab (2022). Here we focused on the implementation of a field-based recommendation system that provides real-time feedback. The recommender system based on question answer pair was is built using machine learning and natural language processing techniques. The work tested various algorithms and settled with a retrieval based model, RetBERT which applies a BERT sentence transformer and has a BLEU score accuracy of 50.8\% limited by the amount of available dataset. Our application tools integrates both online and offline services since farmers come from remote areas where internet is limited. Success in this work will result in a large trial to validate its applicability for use in alleviating the food security problem in Sub-Saharan Africa.