A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks
A sufficient condition that a region be classifiable by a two-layer feedforward neural net (a two-layer perceptron) using threshold activation functions is that either it be a convex polytope or that intersected with the complement of a convex polytope in its interior, or that intersected with the c...
Guardado en:
Autores principales: | , |
---|---|
Publicado: |
2000
|
Materias: | |
Acceso en línea: | https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_10459227_v11_n3_p811_Cabrelli http://hdl.handle.net/20.500.12110/paper_10459227_v11_n3_p811_Cabrelli |
Aporte de: |
id |
paper:paper_10459227_v11_n3_p811_Cabrelli |
---|---|
record_format |
dspace |
spelling |
paper:paper_10459227_v11_n3_p811_Cabrelli2023-06-08T16:01:11Z A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks Cabrelli, Carlos Alberto Molter, Ursula Maria Convex recursive deletion region Neural network Two-layer perceptron Algorithms Backpropagation Computational geometry Matrix algebra Multilayer neural networks Optimization Theorem proving Vectors Convex recursive deletion region Two layer perceptron Feedforward neural networks A sufficient condition that a region be classifiable by a two-layer feedforward neural net (a two-layer perceptron) using threshold activation functions is that either it be a convex polytope or that intersected with the complement of a convex polytope in its interior, or that intersected with the complement of a convex polytope in its interior or . . . recursively. These have been called convex recursive deletion (CoRD) regions. We give a simple algorithm for finding the weights and thresholds in both layers for a feedforward net that implements such a region. The results of this work help in understanding the relationship between the decision region of a perceptron and its corresponding geometry in input space. Our construction extends in a simple way to the case that the decision region is the disjoint union of CoRD regions (requiring three layers). Therefore this work also helps in understanding how many neurons are needed in the second layer of a general three-layer network. In the event that the decision region of a network is known and is the union of CoRD regions, our results enable the calculation of the weights and thresholds of the implementing network directly and rapidly without the need for thousands of backpropagation iterations. © 2000 IEEE. Fil:Cabrelli, C. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales; Argentina. Fil:Molter, U. Universidad de Buenos Aires. Facultad de Ciencias Exactas y Naturales; Argentina. 2000 https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_10459227_v11_n3_p811_Cabrelli http://hdl.handle.net/20.500.12110/paper_10459227_v11_n3_p811_Cabrelli |
institution |
Universidad de Buenos Aires |
institution_str |
I-28 |
repository_str |
R-134 |
collection |
Biblioteca Digital - Facultad de Ciencias Exactas y Naturales (UBA) |
topic |
Convex recursive deletion region Neural network Two-layer perceptron Algorithms Backpropagation Computational geometry Matrix algebra Multilayer neural networks Optimization Theorem proving Vectors Convex recursive deletion region Two layer perceptron Feedforward neural networks |
spellingShingle |
Convex recursive deletion region Neural network Two-layer perceptron Algorithms Backpropagation Computational geometry Matrix algebra Multilayer neural networks Optimization Theorem proving Vectors Convex recursive deletion region Two layer perceptron Feedforward neural networks Cabrelli, Carlos Alberto Molter, Ursula Maria A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks |
topic_facet |
Convex recursive deletion region Neural network Two-layer perceptron Algorithms Backpropagation Computational geometry Matrix algebra Multilayer neural networks Optimization Theorem proving Vectors Convex recursive deletion region Two layer perceptron Feedforward neural networks |
description |
A sufficient condition that a region be classifiable by a two-layer feedforward neural net (a two-layer perceptron) using threshold activation functions is that either it be a convex polytope or that intersected with the complement of a convex polytope in its interior, or that intersected with the complement of a convex polytope in its interior or . . . recursively. These have been called convex recursive deletion (CoRD) regions. We give a simple algorithm for finding the weights and thresholds in both layers for a feedforward net that implements such a region. The results of this work help in understanding the relationship between the decision region of a perceptron and its corresponding geometry in input space. Our construction extends in a simple way to the case that the decision region is the disjoint union of CoRD regions (requiring three layers). Therefore this work also helps in understanding how many neurons are needed in the second layer of a general three-layer network. In the event that the decision region of a network is known and is the union of CoRD regions, our results enable the calculation of the weights and thresholds of the implementing network directly and rapidly without the need for thousands of backpropagation iterations. © 2000 IEEE. |
author |
Cabrelli, Carlos Alberto Molter, Ursula Maria |
author_facet |
Cabrelli, Carlos Alberto Molter, Ursula Maria |
author_sort |
Cabrelli, Carlos Alberto |
title |
A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks |
title_short |
A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks |
title_full |
A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks |
title_fullStr |
A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks |
title_full_unstemmed |
A constructive algorithm to solve "Convex Recursive Deletion" (CoRD) classification problems via two-layer perceptron networks |
title_sort |
constructive algorithm to solve "convex recursive deletion" (cord) classification problems via two-layer perceptron networks |
publishDate |
2000 |
url |
https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_10459227_v11_n3_p811_Cabrelli http://hdl.handle.net/20.500.12110/paper_10459227_v11_n3_p811_Cabrelli |
work_keys_str_mv |
AT cabrellicarlosalberto aconstructivealgorithmtosolveconvexrecursivedeletioncordclassificationproblemsviatwolayerperceptronnetworks AT molterursulamaria aconstructivealgorithmtosolveconvexrecursivedeletioncordclassificationproblemsviatwolayerperceptronnetworks AT cabrellicarlosalberto constructivealgorithmtosolveconvexrecursivedeletioncordclassificationproblemsviatwolayerperceptronnetworks AT molterursulamaria constructivealgorithmtosolveconvexrecursivedeletioncordclassificationproblemsviatwolayerperceptronnetworks |
_version_ |
1768544513466302464 |