International audience

Case-based prediction (CBP) methods do not learn a model of the target decision function but instead perform an inference process that depends on two similarity measures and a reference case base. This paper proposes a strategy, called EnergyCompress, to learn an effective case base by selecting relevant cases from an initial set. Use of EnergyCompress decreases CBP inference time, through case base compression, and also increases prediction performance, for a wide variety of CBP algorithms. EnergyCompress relies on the proposition of a general formulation of the CBP task in the framework of energy-based models, which leads to a new and valuable characterization of the notion of competence in case-based reasoning, in particular at the source case level. Extensive experimental results on 18 benchmark datasets comparing EnergyCompress to 5 reference algorithms for case base maintenance support the benefit of the proposed strategy.

Analogical transfer consists in leveraging a measure of similarity between two situations to predict the amount of similarity between their outcomes. Acquiring a suitable similarity measure for analogical transfer may be difficult, especially when the data is sparse or when the domain knowledge is incomplete. To alleviate this problem, this paper presents a dataset complexity measure that can be used either to select an optimal similarity measure, or if the similarity measure is given, to perform analogical transfer: among the potential outcomes of a new situation, the most plausible is the one which minimizes the dataset complexity.

In machine learning, objects are usually grouped according to similarities found in the objects descriptions. Recent works, however, suggest that representing the differences between object descriptions is also pertinent in many learning tasks. But not much study has been made on how to represent and learn from differences. This paper proposes a qualitative representation of inter-object variations that can be used as input of a learning task. The main idea is to define inter-objects variations as attributes of repetitions of objects, so that machine learning methods will be able to manipulate them in the same way as they manipulate object attributes. The approach is tested on both classification and a numerical value prediction tasks and shows encouraging results.