Theory and scope of exact representation extraction from feed-forward networks

Authors:
Ofer Melnik;Jordan B. Pollack
Affiliations:
DIMACS Center, Rutgers University, Piscataway, NJ, USA;Computer Science Department, Brandeis University, Waltham, MA, USA
Venue:
Cognitive Systems Research
Year:
2002

Citing 4
Cited 2

Extracting rules from neural networks by pruning and hidden-unit splitting

Neural Computation
Rule-extraction by backpropagation of polyhedra

Neural Networks
Symbolic Rule Extraction from the DIMLP Neural Network

Hybrid Neural Systems, revised papers from a workshop
Extracting Provably Correct Rules from Artificial Neural Networks

Extracting Provably Correct Rules from Artificial Neural Networks

Decision Region Connectivity Analysis: A Method for Analyzing High-Dimensional Classifiers

Machine Learning
Gene expression analysis of leukemia samples using visual interpretation of small ensembles: a case study

PRIB'07 Proceedings of the 2nd IAPR international conference on Pattern recognition in bioinformatics

Quantified Score

Hi-index	0.00

Visualization

Abstract

An algorithm to extract representations from feed-forward threshold networks is outlined. The representation is based on polytopic decision regions in the input space - and is exact not an approximation. Using this exact representation we explore scope questions, such as when and where do networks form artifacts, or what can we tell about network generalization from its representation. The exact nature of the algorithm also lends itself to theoretical questions about representation extraction in general, such as what is the relationship between factors such as input dimensionality, number of hidden units, number of hidden layers, and how the network output is interpreted to the potential complexity of the network's function.