Case-Based Cooking with Generic Computer Utensils: Taaable Next Generation

Emmanuelle Gaillard (1, 2) , Jean Lieber (1, 2) , Emmanuel Nauer (1, 2)

1 ORPAILLEUR - Knowledge representation, reasonning
2 LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications

This paper presents the participation of the Taaable team in the 2014 Computer Cooking Contest. The three challenges proposed this year are addressed. The basic challenge is addressed with a new version of the Taaable system, built on Tuuurbine, a generic case-based reasoning system over RDFS. The mixology challenge which requires building recipes only by using a set of available foods is also directly addressed using Tuuurbine conjointly with Revisor, an adaptation engine implementing various revision operators. For the mixology challenge, Revisor is used to compute ingredient substitutions and to adjust the ingredient quantities. The taste score is evaluated as the probability that the adapted cocktail is tasty, depending on the probabilities that the retrieved recipe and the adaptation performed are good. The text of the preparation is adapted using textual substitutions. Finally, the originality challenge addresses reasoning on knowledge built collaboratively by an e-community, taking into account the reliability of knowledge units.

Mots clés: case-based reasoning; belief revision; adaptation of ingredient quantities; adaptation of recipe preparations; knowledge reliability

Domaines: Informatique

Emmanuelle Gaillard, Jean Lieber, Emmanuel Nauer. Case-Based Cooking with Generic Computer Utensils: Taaable Next Generation. Proceedings of the ICCBR 2014 Workshops, David B. Leake and Jean Lieber, Sep 2014, Cork, Ireland. pp.254. ⟨hal-01082683⟩