An incremental learning algorithm for functional approximation

Jacques Bahi (1) , Sylvain Contassot-Vivier (2) , Marc Sauget (1)

1 FEMTO-ST - Franche-Comté Électronique Mécanique, Thermique et Optique - Sciences et Technologies (UMR 6174)
2 ALGORILLE - Algorithms for the Grid

This paper presents an incremental learning algorithm for feed-forward neural networks used as approximators of real world data. This algorithm allows neural networks of limited size to be obtained, providing better performances. The algorithm is compared to two of the main incremental algorithms (Dunkin and cascade correlation) in the respective contexts of synthetic data and of real data consisting of radiation doses in homogeneous environments.

Mots clés: Neural networks; Incremental Learning

Domaines: Calcul parallèle, distribué et partagé [cs.DC]

Jacques Bahi, Sylvain Contassot-Vivier, Marc Sauget. An incremental learning algorithm for functional approximation. Advances in Engineering Software, 2009, 40 (8), pp.725-730. ⟨10.1016/j.advengsoft.2008.12.018⟩. ⟨inria-00430288⟩