TASTEset - Recipe and Food Entities Dataset

Food computing is currently a fast-growing field of research. Web mining and content analysis are also increasingly essential in this field, especially for recognising food entities.However, only a few well-defined tasks still serve as benchmarks for solutions in this area. To bridge this gap, we introduce a new dataset -- called \textit{TASTEset}. In this dataset, Named Entity Recognition (NER) models are expected to find or infer various types of entities helpful in processing recipes, e.g. food products, quantities and their units, names of cooking processes, physical qualities of ingredients, their purpose, and taste. The dataset consists of 1,000 recipes with about 19,500 entities to extract. We share the dataset and the task for entities extraction to encourage progress on more in-depth and complex information extraction from recipes.