Section: New Results

Around the Taaable research project

Participants : Valmi Dufour-Lussier, Emmanuelle Gaillard, Laura Infante Blanco, Florence Le Ber, Jean Lieber, Amedeo Napoli, Emmanuel Nauer.

knowledge representation, description logics, classification-based reasoning, case-based reasoning, belief revision, semantic web

The Taaable project (http://taaable.fr ) has been originally created as a challenger of the Computer Cooking Contest (ICCBR Conference). A candidate to this contest is a system whose goal is to solve cooking problems on the basis of a recipe book (common to all candidates), where each recipe is a shallow XML document with an important plain text part. The size of the recipe book (about 1500 recipes) prevents from a manual indexing of recipes: this indexing is performed using semi-automatic techniques.

Beyond its participation to the CCCs, the Taaable project aims at federating various research themes: case-based reasoning, information retrieval, knowledge acquisition and extraction, knowledge representation, minimal change theory, ontology engineering, semantic wikis, text-mining, etc. Case-based reasoning is used to perform adaptation of recipe to user constraints. The reasoning process uses a cooking domain ontology (especially hierarchies of classes) and adaptation rules. The knowledge base used by the inference engine is encoded within a semantic wiki, which contains the recipes, the domain ontology, and adaptation rules.

The most important original features of this version are:

Modules for computing adaptation knowledge.

Using adaptation knowledge, and especially adaptation rules, is a way to better adapt cooking recipes to user constraints. A previous work for extracting adaptation rules has been performed in 2011  [93] . In this work, variation of ingredients between couple of recipes are mined using closed itemsets extraction. The adaptation rules come from the interpretation of closed itemsets whose items correspond to the ingredients that have to be removed, kept, or added. This approach has been integrated as a wiki extension, providing a collaborative environment in which humans and machines may now collaborate to better acquire adaptation rules [38] . Humans (expert in cooking) may trigger automatic processes (knowledge discovery processes) and and may validate, using a specific user interface, proposition of adaptation rules as adaptation knowledge, which is then added to the knowledge base. In the same way, this environment integrates also the results of a new work on knowledge extraction where specific cooking adaptation rules (i.e. that can be applied to a single recipe) are generalized using close itemsets into generic adaptation rules, to make them usable on other recipes [60] .

A module for acquiring a process semantic representation.

While a process for acquiring cases from recipe preparation texts exists, the results are not perfect. In order for valid case representations to be available in the semantic wiki, a semi-automatic case acquisition tool was created [59] . This tools presents the user with a graphical interface through which it is able to interact with the case acquisition process. In order to limit the effort required, each correction entered by the user is propagated by the tool to the rest of the case representation.

Some other theoretical studies have been carried out that should be applied to some future versions of Taaable:

• The combination of workflows and interval algebras to represent procedural knowledge [55] .

• The revision-based adaptation of cases represented in a qualitative algebra [41] .

• The study of taxonomy merging [39] : several versions of the taxonomies used in Taaable (such as the food hierarchy) can be incoherent one with the others and a merging process is defined in order to obtain a consistent merged taxonomy.

• A continuous knowledge extraction process to ensure the non regression of the reasoning system according to the ontology evolution [50] .