Neural Networks as Artificial Specifications

I. S. W. B. Prasetya; M. A.  Tran

doi:10.1007/978-3-319-99927-2_11

Neural Networks as Artificial Specifications

I. S. W. B. Prasetya^*, M. A. Tran

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

In theory, a neural network can be trained to act as an artificial specification for a program by showing it samples of the programs executions. In practice, the training turns out to be very hard. Programs often operate on discrete domains for which patterns are difficult to discern. Earlier experiments reported too much false positives. This paper revisits an experiment by Vanmali et al. by investigating several aspects that were uninvestigated in the original work: the impact of using different learning modes, aggressiveness levels, and abstraction functions.
The results are quite promising.

Original language	English
Title of host publication	Testing Software and Systems
Subtitle of host publication	30th IFIP WG 6.1 International Conference, ICTSS 2018, Cádiz, Spain, October 1-3, 2018, Proceedings
Editors	Immaculada Medina-Bulo, Mercedes G. Merayo, Robert Hierons
Place of Publication	Cham
Publisher	Springer
Pages	135-141
Number of pages	7
Edition	1
ISBN (Electronic)	978-3-319-99927-2
ISBN (Print)	978-3-319-99926-5
DOIs	https://doi.org/10.1007/978-3-319-99927-2_11
Publication status	Published - 7 Sept 2018

Publication series

Name	Lecture notes in computer science
Publisher	Springer
Volume	11146
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Keywords

Neural network for software testing
Automated oracles

Access to Document

10.1007/978-3-319-99927-2_11

prasetyaFinal published version, 931 KBLicence: Taverne

Cite this

Prasetya, I. S. W. B., & Tran, M. A. (2018). Neural Networks as Artificial Specifications. In I. Medina-Bulo, M. G. Merayo, & R. Hierons (Eds.), Testing Software and Systems: 30th IFIP WG 6.1 International Conference, ICTSS 2018, Cádiz, Spain, October 1-3, 2018, Proceedings (1 ed., pp. 135-141). (Lecture notes in computer science; Vol. 11146). Springer. https://doi.org/10.1007/978-3-319-99927-2_11

@inproceedings{cc1389b47d5842c9ba33207b169039a3,

title = "Neural Networks as Artificial Specifications",

abstract = "In theory, a neural network can be trained to act as an artificial specification for a program by showing it samples of the programs executions. In practice, the training turns out to be very hard. Programs often operate on discrete domains for which patterns are difficult to discern. Earlier experiments reported too much false positives. This paper revisits an experiment by Vanmali et al. by investigating several aspects that were uninvestigated in the original work: the impact of using different learning modes, aggressiveness levels, and abstraction functions.The results are quite promising.",

keywords = "Neural network for software testing, Automated oracles",

author = "Prasetya, {I. S. W. B.} and Tran, {M. A.}",

year = "2018",

month = sep,

day = "7",

doi = "10.1007/978-3-319-99927-2_11",

language = "English",

isbn = "978-3-319-99926-5",

series = "Lecture notes in computer science",

publisher = "Springer",

pages = "135--141",

editor = "Immaculada Medina-Bulo and Merayo, {Mercedes G.} and Robert Hierons",

booktitle = "Testing Software and Systems",

edition = "1",

}

Prasetya, ISWB & Tran, MA 2018, Neural Networks as Artificial Specifications. in I Medina-Bulo, MG Merayo & R Hierons (eds), Testing Software and Systems: 30th IFIP WG 6.1 International Conference, ICTSS 2018, Cádiz, Spain, October 1-3, 2018, Proceedings. 1 edn, Lecture notes in computer science, vol. 11146, Springer, Cham, pp. 135-141. https://doi.org/10.1007/978-3-319-99927-2_11

Neural Networks as Artificial Specifications. / Prasetya, I. S. W. B.; Tran, M. A. .
Testing Software and Systems: 30th IFIP WG 6.1 International Conference, ICTSS 2018, Cádiz, Spain, October 1-3, 2018, Proceedings. ed. / Immaculada Medina-Bulo; Mercedes G. Merayo; Robert Hierons. 1. ed. Cham: Springer, 2018. p. 135-141 (Lecture notes in computer science; Vol. 11146).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Neural Networks as Artificial Specifications

AU - Prasetya, I. S. W. B.

AU - Tran, M. A.

PY - 2018/9/7

Y1 - 2018/9/7

N2 - In theory, a neural network can be trained to act as an artificial specification for a program by showing it samples of the programs executions. In practice, the training turns out to be very hard. Programs often operate on discrete domains for which patterns are difficult to discern. Earlier experiments reported too much false positives. This paper revisits an experiment by Vanmali et al. by investigating several aspects that were uninvestigated in the original work: the impact of using different learning modes, aggressiveness levels, and abstraction functions.The results are quite promising.

AB - In theory, a neural network can be trained to act as an artificial specification for a program by showing it samples of the programs executions. In practice, the training turns out to be very hard. Programs often operate on discrete domains for which patterns are difficult to discern. Earlier experiments reported too much false positives. This paper revisits an experiment by Vanmali et al. by investigating several aspects that were uninvestigated in the original work: the impact of using different learning modes, aggressiveness levels, and abstraction functions.The results are quite promising.

KW - Neural network for software testing

KW - Automated oracles

U2 - 10.1007/978-3-319-99927-2_11

DO - 10.1007/978-3-319-99927-2_11

M3 - Conference contribution

SN - 978-3-319-99926-5

T3 - Lecture notes in computer science

SP - 135

EP - 141

BT - Testing Software and Systems

A2 - Medina-Bulo, Immaculada

A2 - Merayo, Mercedes G.

A2 - Hierons, Robert

PB - Springer

CY - Cham

ER -

Neural Networks as Artificial Specifications

Abstract

Publication series

Keywords

Access to Document

Fingerprint

Cite this