EN FR
EN FR


Section: New Results

Reasoning with Imperfect Information and Priorities

Participants : Abdallah Arioua, Patrice Buche, Madalina Croitoru, Jérôme Fortin, Nouredine Tamani, Rallou Thomopoulos.

This year, we mainly explored the use of argumentation frameworks in practical applications. Indeed, we have been involved in three main projects that have employed argumentation techniques. The projects were all in the context of agronomy where the nature of the problem studies fits well the use of argumentation: (1) the knowledge bases considered to model the domain are inconsistent, (2) the reasoning / decision making has to take into account the inconsistency, (3) the end user is a non-computing expert thus explanation facilities are needed.

We enumerate below the three projects and explain our results:

  • Bread Making project (financed by the Food and Bioproducts department at INRA) investigates the possibility of using wholemeal flour in bread as opposed to classic white flour. The main theoretical result that we exploited here was the instantiation work into the existential rule framework done with Srdjan Vesic [43] . We used reverse engineering and the subsequent logic-based argumentation in order to provide the experts with a cartography of possible pros and cons of using one type of flour vs the other.

    Ecological Informatics 2015 [18]

  • EcoBioCap (FP7 EU project led by INRA Montpellier, see Section  9.1 ) investigates the conception of biodegradable packaging for fruits and vegetables. The main theoretical result used here concerns the fuzzy aspects of argumentation but the modeling of the problem using argumentation and subsequent argument elicitation was also a very challenging process.

    Computers and Electronics in Agriculture 2015 (2 papers) [14] [17]

  • DURDUR (ANR project led by INRA Montpellier, see Section  9.1 ) investigates the technological itineraries to grow durum wheat for subsequent pasta and couscous making. This ongoing project investigates the use of argumentation for explanation facilities.

    Initial results have been published in SUM 2015[19] and DEXA 2015 [20] .