Abstract
In this chapter, we tackle the problem of extracting symbolic knowledge from trained neural networks; that is, the problem of finding “logical representations” for such networks. We present a new method of extraction that captures nonmonotonic rules encoded in the network and show that such a method is sound. The ideas presented here comply with the concept that, in machine learning, Feature Construction should be decoupled from Model Construction [KLF01]. Part I of this book has dealt with feature construction. Part II will deal with model construction.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2002 Springer-Verlag London
About this chapter
Cite this chapter
d’Avila Garcez, A.S., Broda, K.B., Gabbay, D.M. (2002). Knowledge Extraction from Trained Networks. In: Neural-Symbolic Learning Systems. Perspectives in Neural Computing. Springer, London. https://doi.org/10.1007/978-1-4471-0211-3_5
Download citation
DOI: https://doi.org/10.1007/978-1-4471-0211-3_5
Publisher Name: Springer, London
Print ISBN: 978-1-85233-512-0
Online ISBN: 978-1-4471-0211-3
eBook Packages: Springer Book Archive