This perspective article, co-authored by several Table members including Christian Reynolds, discusses the possibilities and challenges of using natural language processing and artificial intelligence to assess the nutritional and sustainability characteristics of recipes. The authors conclude that an interdisciplinary approach is necessary, combining artificial intelligence with expert domain knowledge.
Challenges identified by the paper include:
- Digital recipe datasets are often not in the same format and may be proprietary.
- Nutrition datasets for different regions are not always directly comparable as they use different codes and hierarchies.
- Sustainability datasets exist for food ingredients (as opposed to entire recipes), but while these do usually follow standards, the scope , functional units and assumptions behind data points are often not documented clearly.
- Datasets on recipes, nutrition and sustainability are rarely directly linked and often use differing terminology.
- Natural language processing can be hampered by factors such as errors in optical character recognition, the need for contextual knowledge (e.g. knowing that teacups have different sizes in Europe and North America), and the fact that not all recipes specify how many portions they are intended to produce.
In this paper, we discuss the use of natural language processing and artificial intelligence to analyse nutritional and sustainability aspects of recipes and food. We present the state-of-the-art and some use cases, followed by a discussion of challenges. Our perspective on addressing these is that while they typically have a technical nature, they nevertheless require an interdisciplinary approach combining natural language processing and artificial intelligence with expert domain knowledge to create practical tools and comprehensive analysis for the food domain.
Van Erp, M., Reynolds, C., Maynard, D., Starke, A., Ibáñez Martín, R., Andres, F., Leite, M.C.A., Alvarez de Toledo, D., Schmidt Rivera, X., Trattner, C., Brewer, S., Martins, C. A., Kluczkovski, A., Frankowska, A., Bridle, S., Levy, R. B., Rauber, F., da Silva, J. T. and Bosma, U. (2021). Using Natural Language Processing and Artificial Intelligence to explore the nutrition and sustainability of recipes and food. Frontiers in Artificial Intelligence, 3(621577).