2025 September 3

Trends review publishes our perspective on advancing AI in food science




The article, titled “Domain knowledge, just evaluation, and robust data standards are required to advance AI in food science” (DOI: 10.1016/j.tifs.2025.105272), highlights the urgent need for stronger foundations in applying artificial intelligence (AI) to food science.

 

Key Findings

The perspective identifies several critical gaps in current food AI research, including fragmented datasets, limited reproducibility, and insufficient benchmarking. The authors propose five strategic directions to accelerate progress:

  • Integration of domain knowledge into AI model development

  • Transparent and reproducible workflows

  • Systematic benchmarking and just evaluation

  • Practical validation through real-world applications

  • Robust data standards and open infrastructure to enable collaboration

By addressing these gaps, the paper argues, AI can more reliably support advances in nutrition, food safety, sustainability, and flavor science.

 

A Global Collaboration

The work is the result of an international collaboration between researchers from Singapore, Ireland, Switzerland, the United States, China, and Canada. Alongside Dr. Zhang, co-authors include Meihui Liu (NUS & Bezos Centre for Sustainable Proteins), Zhaoshuo Yu and Hanlin Xu (University College Dublin), Stephan Pfister (ETH Zurich), Giulia Menichetti (Northeastern University & Harvard Medical School), Xingran Kou (Shanghai Institute of Technology), Jinlin Zhu and Daming Fan (Jiangnan University), and Pingfan Rao (International Union of Food Science and Technology).

 

The study was supported by the Ministry of Education, Singapore (Academic Research Fund Tier 1) and the NUS Start-Up Grant.

 

https://www.sciencedirect.com/science/article/abs/pii/S092422442500408X