Researchers at the Technical University of Munich developed an AI method combining a protein language model with a Graph Convolutional Network to predict bitterness in peptides formed during food production. The model successfully predicted the taste of newly designed peptides, with 25 of 31 tested candidates correctly classified, enabling better control of flavor in fermented foods and plant-based proteins.
Vertical farming—growing plants in stacked layers using hydroponics, aeroponics, or other soil-less methods—is reshaping food production in response to climate change, water scarcity, and urbanization. These systems can reduce water use by 70–95%, shorten supply chains, and allow food production in cities using repurposed buildings, while raising questions about the sensory and cultural dimensions of agriculture.