Rules

To participate, you have:

• to make a statement of interest, now
• to submit a technical paper describing your contribution for the CCC 2014 (The paper length is up to 8 pages if only 1 challenge is addressed and 4 pages more per addressed challenge)
• to demonstrate your system (or approach in case of the originality challenge) in a 20 min presentation during the CCC workshop (at least one member of each team has to register for CCC 2014)

In addition, you may:

• send in a set of up to 200 recipes with a general focus and up to 200 recipes on cocktails in CCC XML format or in WikiTaable until the deadline for case base extension.
• provide a query which will be used for system evaluation of all teams

A running system must be available for the system evaluation (online or on a laptop) during the contest except you contribute to the originality challenge only.

The system must be queryable with a query containing one or more (but at least one) desired ingredient(s) and may contain undesired ingredient(s). The ingredient names are taken from the WikiTaable ontology in the version at the publication time of the final case bases.

The system must use the recipes from the final case base(s) provided by the CCC organizing team. Additional recipes may be used in the background. However, only original recipes from the provided case base(s) are eligible as a starting point for adaptation.

The system must return recipes satisfying the query. For a blind evaluation (basic challenge and mixology challenge), the systems must return results under the XML format given by the following DTD :

<!ELEMENT results (query, result*) >
<!ELEMENT result  (taste_score, original_recipe, final_recipe) >
<!ATTLIST result  rank     CDATA  	#IMPLIED >
<!ELEMENT original_recipe ... see XML example ... >
<!ELEMENT final_recipe ... see XML example ... >

For example :

<results>
<query>cake_dish pear</query>
<result rank=”...”>
	<taste_score>... 0 to 5 ...</taste_score>
	<original_recipe>... in  XML format ...</original_recipe>
	<final_recipe>... in  XML format ...</final_recipe>
     </result>
</results>

The query element contains the query. The taste_score element contains one of the following numbers 0 (horrible), 1 (consumable), 2 (quite good), 3 (tasty), 4 (very tasty), and 5 (fantastic). The orginal_recipe element contains the complete recipe that has been retrieved from the given case base before adaptation (in the CCC XML format for recipes, see http://wikitaaable.loria.fr/rdf/ccc_cocktails.xml for illustration). The final_recipe element contains the complete recipe after adaptation (in the CCC XML format for recipes).

Two initial sets of recipes are provided by the CCC organizing team (see resources) that may be extended by participants as follows:

• Send in up to two hundred recipes per challenge to the CCC organizing team until the recipe submission deadline. The format is either a file with the cases in the CCC XML format sent by mail to ccc@liris.cnrs.fr or you may type in the recipes directly into http://wikitaaable.loria.fr.
• The CCC organizing team will include all submitted recipes into WikiTaaable and provide an export of the final case bases into the CCC XML format until the deadline for publication of the final recipe bases.

The evaluation will be based:

• on the technical paper describing your contribution, for each challenge. Three members of the program commitee will evaluate your work according to the scientific significance (0..7 points), the presentation of the paper (0..2 points), and a brief result section (if possible, 0..1 points).

• on the evaluation of the system by a jury, for the basic and the mixology challenge. All systems will be tested on the same set of queries. Reviewer pairs (1 jury member, not involved in any system + 1 volunteer) will test, in a restrictive time, each system on 1 query per addressed challenge assessing the estimated culinary quality (0 (horrible), 2 (consumable), 4 (quite good), 6 (tasty), 8 (very tasty), or 10 (fantastic)).

• on evaluation of the system by the public, only for the mixology challenge. Volunteers from the public will fill out charts to assess the reality factor (0..3 points), the culinary quality (0 (horrible), 1 (consumable), 2 (quite good), 3 (tasty), 4 (very tasty), or 5 (fantastic)) and the self-awareness factor (0 (more than one point difference between taste_score and public assessment), 1 (one point difference) or 2 points (exact self-assessment)). The maximum reality factor is diminished by one for each ingredient that is not in the restricted set of given ingredients but occurs in the recipe.

• the basic challenge:
  • 10 points will be attributed for the technical paper
  • 10 points will be attributed for the system evaluation (average of scores given by reviewer pairs)
• the mixology challenge:
  • 10 points will be attributed for the technical paper
  • 10 points will be attributed after testing the cocktails (average of scores given by the public)
• the crowd favourite of the mixology challenge:
  • 10 points will be attributed after testing the cocktails (average of scores given by the public)
• the originality challenge:
  • 10 points will be attributed for the technical paper

Technical papers have to describe your contribution(s). Papers addressing system challenges (basic, mixology challenge, or both) may include a brief section on sample results. Papers have to be in Springer format. Paper size is max 8 pages in case you attend one challenge only and max 12 pages in case you attend the basic and the mixology challenge. Teams addressing a system challenge (basic, mixology challenge) might in addition submit a contribution to the originality challenge if it is on a different topic than the system challenge(s).

Submissions have to be made via Easychair (choose the “Computer Cooking Contest” track).

Any individual or group (e.g., academic, industry, or government) is invited to participate. At least one person per finalist must register for the conference, demonstrate the system, and give a talk in the CCC workshop.

We do not make any restrictions on the technology to be used:

• The queries might be transformed manually to an arbitrary input format (structural/formula-based, conversational, text-based) to be processed by the software.
• The representation of the recipes can be chosen arbitrarily.
• The technology to be used for selecting and modifying recipes is absolutely open.
• You can use any kind of supporting software tools; creativity is encouraged!