Skip to Main content Skip to Navigation

Interactive and Opportunistic Knowledge Acquisition in Case-Based Reasoning

Amélie Cordier 1
1 SILEX - Supporting Interaction and Learning by Experience
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : As a young discipline at the junction of computer science, artificial intelligence and cognitive sciences, knowledge engineering aims at modelling knowledge of a specific domain to operationalise them in a computer system. To this end, it offers theoretical tools, models and empirical methodologies to support knowledge sharing between the user and the system.
The work developed here is related to knowledge engineering of a particular type of system: case-based reasoning systems (CBR). A CBR system assists a user in his problem solving task by retrieving a previous successful problem solving experience and by adapting it to the current situation. In this work, we are mainly interested in the interacting system "user - CBR tool".
The main research question we address here can be formulated as: what methods and tools have to be developed to support knowledge acquisition in the learning system "user - CBR tool". This issue raises the question of the knowledge of the reasoning process and leads to an analysis at the knowledge level of CBR systems. Another part of the analysis aims at studying the interactions between the user and the CBR tool during the problem solving phases. These aspects are studied at several levels in the different contributions presented in this thesis.
Our different experiences and experiments lead us to propose, as a first contribution, a formalisation at general level of interactive knowledge learning in CBR (FIKA). This formalisa- tion relies on the reasoning failures which, as they allow to highlight the gaps in the available knowledge, are used to guide the learning process. Two extensions of this general model have been proposed: IAKA and FRAKAS.
IAKA refines the principles proposed in FIKA to permit their immediate implementation in a particular type of system where knowledge can be represented according to a given model (cases and adaptation knowledge in the form of adaptation operators). These principles have been implemented and experimented with in an application developed exclusively for this purpose. FRAKAS proposes similar methods and tools for another type of system where domain knowledge is used to guide adaptation. As for these principles, they have been implemented in a prototype inspired by a real world application.
We have conducted a study of strengths and limits of FRAKAS and IAKA and we have investigated possible ways to combine them. A first practical implementation has been made in a CBR application allowing the adaptation of cooking recipes, the project TAAABLE.
Document type :
Complete list of metadata

Cited literature [170 references]  Display  Hide  Download
Contributor : Amélie Cordier Connect in order to contact the contributor
Submitted on : Wednesday, February 25, 2009 - 11:28:03 PM
Last modification on : Tuesday, June 1, 2021 - 2:08:08 PM
Long-term archiving on: : Friday, October 12, 2012 - 12:31:11 PM


  • HAL Id : tel-00364368, version 1


Amélie Cordier. Interactive and Opportunistic Knowledge Acquisition in Case-Based Reasoning. Human-Computer Interaction [cs.HC]. Université Claude Bernard - Lyon I, 2008. English. ⟨NNT : 2008LYO10224⟩. ⟨tel-00364368⟩



Record views


Files downloads