M. Egmont-Petersen. "An approach for generating explanations in neural networks," In: J. Brender & P. McNair (Eds.), Proceedings for the KAVAS Workshop on Knowledge Acquisition, Visualization, and Assessment, Computer Resources International, Copenhagen, pp. 131-136, 1990.
 
 

Abstract:

This paper gives a brief introduction to the domain of the law used as an example case in an experiment with a neural network. It is explained how the sources of law were formalized, a method to build an explanation around the network's decision is being lined out. Finally a few preliminary conclusions and experiences are being summed up.

Reprints, please contact me: michael@cs.uu.nl