Inductive Learning from State Transitions over Continuous Domains

Tony Ribeiro (1, 2) , Sophie Tourret (3) , Maxime Folschette (1, 2, 4) , Morgan Magnin (1, 2) , Domenico Borzacchiello (5) , Francisco Chinesta (5) , Olivier Roux (1, 2) , Katsumi Inoue (3)

1 LS2N - équipe MéForBio - Méthodes Formelles pour la Bioinformatique
2 LS2N - Laboratoire des Sciences du Numérique de Nantes
3 NII - National Institute of Informatics
4 Irset - Institut de recherche en santé, environnement et travail
5 ICI - Institut de Calcul Intensif

Learning from interpretation transition (LFIT) automatically constructs a model of the dynamics of a system from the observation of its state transitions. So far, the systems that LFIT handles are restricted to discrete variables or suppose a discretization of continuous data. However, when working with real data, the discretization choices are critical for the quality of the model learned by LFIT. In this paper, we focus on a method that learns the dynamics of the system directly from continuous time-series data. For this purpose, we propose a modelling of continuous dynamics by logic programs composed of rules whose conditions and conclusions represent continuums of values.

Mots clés: Continuous logi...; Dynamical syste...; Inductive logic...; Learning from i...; dynamical systems; learning from interpretation transition; Continuous Logic Programming; Inductive Logic Programming

Domaines: Modélisation et simulation; Bio-informatique [q-bio.QM]

Tony Ribeiro, Sophie Tourret, Maxime Folschette, Morgan Magnin, Domenico Borzacchiello, et al.. Inductive Learning from State Transitions over Continuous Domains. The 27th International Conference on Inductive Logic Programming (ILP 2017), Nicolas Lachiche; Christel Vrain, Sep 2017, Orléans, France. pp.124-139, ⟨10.1007/978-3-319-78090-0_9⟩. ⟨hal-01655644v1⟩