The Neural Network is challenged by a new competitor: The Information Continuum. Signals are not bound any longer to the edges of a network. Instead, they are drifting in a continuous space. The whole model is built in a continuous structure, which allows adjusting every parameter in small steps. If powerful training methods are developed then the Information Continuum can become important in all areas reserved for Neural Networks until now.
Print ISSN: 1611-2776
Volume: 41, 06/1999
Pages: 25