Neural plasma

Daniel Berrar, Werner Dubitzky

Research output: Chapter in Book/Report/Conference proceedingChapter

1 Citation (Scopus)

Abstract

This paper presents a novel type of artificial neural network, called neural plasma, which is tailored for classification tasks involving few observations with a large number of variables. Neural plasma learns to adapt its classification confidence by generating artificial training data as a function of its confidence in previous decisions. In contrast to multilayer perceptrons and similar techniques, which are inspired by topological and operational aspects of biological neural networks, neural plasma is motivated by aspects of high-level behavior and reasoning in the presence of uncertainty. The basic principles of the proposed model apply to other supervised learning algorithms that provide explicit classification confidence values. The empirical evaluation of this new technique is based on benchmarking experiments involving data sets from biotechnology that are characterized by the small-n-large-p problem. The presented study exposes a comprehensive methodology and is seen as a first step in exploring different aspects of this methodology.

Original languageEnglish
Title of host publicationIFIP International Federation for Information Processing
Pages159-168
Number of pages10
Volume217
DOIs
Publication statusPublished - 2006
Externally publishedYes

Publication series

NameIFIP International Federation for Information Processing
Volume217
ISSN (Print)15715736

Fingerprint

Plasma
Confidence
Methodology
Artificial neural network
Empirical evaluation
Uncertainty
Biotechnology
Neural networks
Benchmarking
Experiment
Learning algorithm

ASJC Scopus subject areas

  • Information Systems and Management

Cite this

Berrar, D., & Dubitzky, W. (2006). Neural plasma. In IFIP International Federation for Information Processing (Vol. 217, pp. 159-168). (IFIP International Federation for Information Processing; Vol. 217). https://doi.org/10.1007/978-0-387-34747-9_17

Neural plasma. / Berrar, Daniel; Dubitzky, Werner.

IFIP International Federation for Information Processing. Vol. 217 2006. p. 159-168 (IFIP International Federation for Information Processing; Vol. 217).

Research output: Chapter in Book/Report/Conference proceedingChapter

Berrar, D & Dubitzky, W 2006, Neural plasma. in IFIP International Federation for Information Processing. vol. 217, IFIP International Federation for Information Processing, vol. 217, pp. 159-168. https://doi.org/10.1007/978-0-387-34747-9_17
Berrar D, Dubitzky W. Neural plasma. In IFIP International Federation for Information Processing. Vol. 217. 2006. p. 159-168. (IFIP International Federation for Information Processing). https://doi.org/10.1007/978-0-387-34747-9_17
Berrar, Daniel ; Dubitzky, Werner. / Neural plasma. IFIP International Federation for Information Processing. Vol. 217 2006. pp. 159-168 (IFIP International Federation for Information Processing).
@inbook{ca9881ca6f96412eb194ad87b579e97f,
title = "Neural plasma",
abstract = "This paper presents a novel type of artificial neural network, called neural plasma, which is tailored for classification tasks involving few observations with a large number of variables. Neural plasma learns to adapt its classification confidence by generating artificial training data as a function of its confidence in previous decisions. In contrast to multilayer perceptrons and similar techniques, which are inspired by topological and operational aspects of biological neural networks, neural plasma is motivated by aspects of high-level behavior and reasoning in the presence of uncertainty. The basic principles of the proposed model apply to other supervised learning algorithms that provide explicit classification confidence values. The empirical evaluation of this new technique is based on benchmarking experiments involving data sets from biotechnology that are characterized by the small-n-large-p problem. The presented study exposes a comprehensive methodology and is seen as a first step in exploring different aspects of this methodology.",
author = "Daniel Berrar and Werner Dubitzky",
year = "2006",
doi = "10.1007/978-0-387-34747-9_17",
language = "English",
isbn = "0387346554",
volume = "217",
series = "IFIP International Federation for Information Processing",
pages = "159--168",
booktitle = "IFIP International Federation for Information Processing",

}

TY - CHAP

T1 - Neural plasma

AU - Berrar, Daniel

AU - Dubitzky, Werner

PY - 2006

Y1 - 2006

N2 - This paper presents a novel type of artificial neural network, called neural plasma, which is tailored for classification tasks involving few observations with a large number of variables. Neural plasma learns to adapt its classification confidence by generating artificial training data as a function of its confidence in previous decisions. In contrast to multilayer perceptrons and similar techniques, which are inspired by topological and operational aspects of biological neural networks, neural plasma is motivated by aspects of high-level behavior and reasoning in the presence of uncertainty. The basic principles of the proposed model apply to other supervised learning algorithms that provide explicit classification confidence values. The empirical evaluation of this new technique is based on benchmarking experiments involving data sets from biotechnology that are characterized by the small-n-large-p problem. The presented study exposes a comprehensive methodology and is seen as a first step in exploring different aspects of this methodology.

AB - This paper presents a novel type of artificial neural network, called neural plasma, which is tailored for classification tasks involving few observations with a large number of variables. Neural plasma learns to adapt its classification confidence by generating artificial training data as a function of its confidence in previous decisions. In contrast to multilayer perceptrons and similar techniques, which are inspired by topological and operational aspects of biological neural networks, neural plasma is motivated by aspects of high-level behavior and reasoning in the presence of uncertainty. The basic principles of the proposed model apply to other supervised learning algorithms that provide explicit classification confidence values. The empirical evaluation of this new technique is based on benchmarking experiments involving data sets from biotechnology that are characterized by the small-n-large-p problem. The presented study exposes a comprehensive methodology and is seen as a first step in exploring different aspects of this methodology.

UR - http://www.scopus.com/inward/record.url?scp=33845533969&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33845533969&partnerID=8YFLogxK

U2 - 10.1007/978-0-387-34747-9_17

DO - 10.1007/978-0-387-34747-9_17

M3 - Chapter

AN - SCOPUS:33845533969

SN - 0387346554

SN - 9780387346557

VL - 217

T3 - IFIP International Federation for Information Processing

SP - 159

EP - 168

BT - IFIP International Federation for Information Processing

ER -