In this paper, a new hybrid adaptation model for cancer diagnosis has been developed. It combines transformational and hierarchical adaptation techniques with artificial neural ne...
This paper proposes a new connectionist approach to numeric law discovery; i.e., neural networks (law-candidates) are trained by using a newly invented second-order learning algor...
Abstract. Reservoir computing approaches have been successfully applied to a variety of tasks. An inherent problem of these approaches, is, however, their variation in performance ...
Segmentation of speech signals is a crucial task in many types of speech analysis. We present a novel approach at segmentation on a syllable level, using a Bidirectional Long-Shor...
Christian Landsiedel, Jens Edlund, Florian Eyben, ...
This paper proposes a novel constructive learning algorithm for a competitive neural network. The proposed algorithm is developed by taking ideas from the immune system and demons...
Helder Knidel, Leandro Nunes de Castro, Fernando J...