Preference moore machines for neural fuzzy integration

This paper describes multidimensional neural preference classes and preference Moore ma-chines as a principle for integrating different neural and/or symbolic knowledge sources. We relate neural preferences to multidimensional fuzzy set representations. Furthermore, we in-troduce neural preference Moore machines and relate traditional symbolic transducers with simple recurrent networks by using neural pref-erence Moore machines. Finally, we demon-strate how the concepts of preference classes and preference Moore machines can be used to integrate knowledge from different neural and/or symbolic machines. We argue that our new concepts for preference Moore machines contribute a new potential approach towards general principles of neural symbolic integra-tion.