Yuchen Ma, Ling Fu, Fatao He, Di Wu, Y. S. Lin, Dejian Huang, Caili Fu, Caoxing Huang
Food by-products are rich in nutrients but are often discarded, causing resource waste and environmental burden. Traditional additive manufacturing (AM) struggles with these materials due to inconsistent rheology, unstable transformations, and complex multiaxis operations. This review explores integrating machine learning (ML) with multidimensional food printing (FP) to valorize by-products. It highlights the use of animal-, plant-, and oilseed-based by-products in 3D printing and their functional transformation in 4D printing. ML enhances the AM pipeline by predicting rheology, optimizing formulations, and enabling real-time process control. It supports adaptive printing, deformation prediction, and closed-loop path adjustments for improved product quality. While 5D/6D printing remains emerging, ML can drive complex structure construction. Key challenges include limited data, poor model transferability, and high computation costs. Future integration with IoT and cloud platforms may enable autonomous, scalable, zero-waste food manufacturing. This ML-driven approach fosters sustainable production and human-AI collaboration.