Utilizing artificial intelligence to enhance sustainable high-quality food production: Current state, challenges, and future directions.

Abstract

Food science today has undergone a profound change with the increasing impact of analyzing heterogeneous multimodal data with artificial intelligence (AI). The adoption of ubiquitous sensors, cameras, and mobile apps in agriculture, food production, food processing, transportation, and personal nutrition monitoring can record vast multimodal data, such as tabular, image, text, omics, sensors, social media, transactions and trade data. The analysis of such multimodal data is usually achieved by AI-based food science and nutrition systems (AIFNS) through the following pipeline: data preprocessing, segmentation, dimensionality reduction, classification, or clustering. When multiple data sources are available for the same prediction task, data fusion strategies can also be employed. To ensure that this new generation of intelligent systems for food science and nutrition can better serve both industry and society, further research is needed to ensure that they operate ethically, transparently, inclusively, and safely across borders.