Towards Physics-Informed Artificial General Intelligence for Food
Please login to view abstract download link
Food is integral to human and planetary health, and alternative proteins offer a promising route to reduce the health and environmental impacts of animal-based meat consumption [3]. From a mechanics point, foods are complex materials that exhibit a finite poro-visco-elastic behavior [1]; yet, their constitutive properties remain poorly understood. Existing empirical descriptors provide limited mechanistic insight into the relationship between food mechanics and taste. Physics-based machine learning holds promise to address this limitation. Here we show how constitutive neural networks empower the automated discovery of constitutive models that quantitatively compare fungi-, plant-, and animal-based foods across multiple deformation modes. We use systematic tension, compression, and shear tests [4], rheological measurements [5], and texture profile analysis [2] to train the networks and discover interpretable and generalizable constitutive models for a wide variety of foods. We use these models to quantify which mechanical features alternative meat products can already successfully replicate and where critical discrepancies to animal meats persist. By integrating our mechanical analysis with sensory studies across dining halls and restaurant settings, we establish quantitative links between mechanics, texture, and taste [4]. Our work outlines a path toward automated science for food design [3], in which autonomous laboratories close the loop between experiment, machine learning based data analysis, automated model discovery, and sensory feedback. This vision highlights opportunities for artificial general intelligence—scientifically grounded in physics and mechanics—to accelerate sustainable food innovation towards human and planetary health. REFERENCES [1] Boes B, et al. The mechanics and physics of tofu: Understanding hydrated soft solids through feature networks. bioRxiv. doi:10.64898/2025.12.10.693552. [2] Dunne RA, et al. Texture profile analysis and rheology of plant-based and animal meat; Food Res Int. 2025; 115876. [3] Kuhl E. AI for Food: Accelerating and democratizing discovery and innovation. npj Science of Food. 2025; 9: 82. [4] St. Pierre SR, et al.The mechanical and sensory signature of plant-based and animal meat. npj Science of Food. 2024; 8: 94. [5] Vervenne T, St Pierre SR, Famaey N, Kuhl E. Probing mycelium mechanics and taste: The moist and fibrous signature of fungi steak. Acta Biomat. 2025; 202: 341-351.
