WP1
Title: Data and knowledge acquisition to build gold and silver standard corpora of recipes and their ingredients with their roles in different contexts.
Objectives:
To gather and integrate domain ontologies and other relevant knowledge resources (such as knowledge graphs) in food domain
To annotate entities and relations in a recipe corpus with concepts from a knowledge graph
To train embeddings of knowledge graphs and recipe structured texts
WP2
Title: Development of a machine learning-based models for recommending substitutable ingredients and their roles.
Objectives:
To develop a deep learning-based model for identifying substitutable ingredients
To develop a deep learning-based model for predicting the role of an ingredient in a given context, e.g. a special diet, or a particular dish
WP3
Title: Development of a logic-based solution for selecting target ingredients and their plausible substitutes
Objectives:
To develop a method availing of logical reasoning to prune and select candidate substitutes based on constraints.
To develop an approach to generate explanations on target ingredients and its replacement.
