AgriUXE: Integrating Explainable AI and Multimodal Data for Smart Agriculture

The increasing demand for sustainable and efficient agricultural practices has driven the adoption of digital technologies in the sector. Despite the persistent lack of transparency in machine learning models, farmers often perceive smart agriculture solutions as complex and of limited value. To address this challenge, we present AgriUXE (Agricultural eXperience Enhanced through eXplainability), a digital platform that combines multimodal data with explainable artifical intelligence (XAI) techniques to provide user-adaptive explanations to drive decision-making processes in smart farming solutions. Given data collected from multimodal sources, such as Internet-of-Things (IoT) sensors and remote sensing, the platform intends to generate comprehensible insights and predictions tailored to the needs of different farm stakeholders. AgriUXE has the potential to integrate with various applications, including mixed reality and digital twin solutions, showcasing its ability to improve the overall farming experience through improved transparency, and decision-making. Additionally, this paper details the functionalities of AgriDash, a practical application for enhanced viticulture data collection, prediction, and visualization.