O uso de Linguagem Natural Controlada para Representacao de Conhecimento por Usuarios Finais

On limited fan-in optimal neural networks

Sergio Roberto P. da SilvaJosiane Mechiori Pinheiro

Valeriu BeiuSorin DraghiciHanna E. Makaruk

This paper proposes the use of controlled natural language as a way to bring domain experts to the first place in the process of acquisition and refinement of knowledge. The proposed language gives value to communicative aspects of language, using linguistics mechanisms such as quantifiers, pronouns, anaphors and ellipses to strengthen the textual cohesion and, therefore, to make the knowledge expression by the domain experts easier. This paper analyses the influence of limited fan-in on the size and VLSI optimality of highly interconnected nets. Two different approaches show that VLSI- and size-optimal discrete neural networks can be obtained for small fan-in values. They have applications to hardware implementations of neural networks. The first approach is based on implementing a certain sub-class of Boolean functions, F/sub n,m/ functions. We show that this class of functions can be implemented in VLSI-optimal (i.e., minimising AT/sup 2/) neural networks of small constant fan-ins. The second approach is based on implementing Boolean functions for which the classical Shannon's decomposition can be used. Such a solution has already been used by Alon-Bruck (1991) to prove bounds on neural networks with fan-ins limited to 2. We generalise the result presented there to arbitrary fan-in, and prove that the size is minimised by small fan-in values, while relative minimum size solutions can be obtained for fan-ins strictly lower than linear. Finally, a size-optimal neural network having small constant fan-ins is suggested for F/sub n,m/ functions.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: