Teaching GP to program like a human software developer: Using Perplexity Pressure to Guide Program Synthesis Approaches

Publication type:
Conference
Metadata:
Autoren
Autoren-URL
https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=fis-test-1&SrcAuth=WosAPI&KeyUT=WOS:000523218400125&DestLinkType=FullRecord&DestApp=WOS_CPL
DOI
10.1145/3321707.3321738
Externe Identifier
  • Clarivate Analytics Document Solution ID: BO7CJ
Zeitschrift
PROCEEDINGS OF THE 2019 GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE (GECCO'19)
Schlüsselwörter
  • Genetic programming
  • Grammatical evolution
  • Software synthesis
  • Genetic improvement
  • Mining software repositories
  • Language models
Paginierung
1065 - 1074
Datum der Veröffentlichung
2019
Status
Published
Titel
Teaching GP to Program Like a Human Software Developer Using Perplexity Pressure to Guide Program Synthesis Approaches

Data source: Web of Science (Lite)

Other metadata sources:
Autoren
DOI
10.1145/3321707.3321738
Zeitschrift
Proceedings of the Genetic and Evolutionary Computation Conference
Name of conference
GECCO '19: Genetic and Evolutionary Computation Conference
Online publication date
2019
Datum der Veröffentlichung
2019
Status
Published
Herausgeber
ACM
Herausgeber URL
http://dx.doi.org/10.1145/3321707.3321738
Datum der Datenerfassung
2023
Titel
Teaching GP to program like a human software developer

Data source: Crossref

Autoren
Editoren
  • Anne Auger
  • Thomas Stützle
ISBN-13
978-1-4503-6111-8
Zeitschrift
GECCO
Paginierung
1065 - 1074
Datum der Veröffentlichung
2019
Herausgeber
ACM
Herausgeber URL
https://doi.org/10.1145/3321707
Titel
Teaching GP to program like a human software developer: using perplexity pressure to guide program synthesis approaches.

Data source: DBLP

Abstract
Program synthesis is one of the relevant applications of GP with a strong impact on new fields such as genetic improvement. In order for synthesized code to be used in real-world software, the structure of the programs created by GP must be maintainable. We can teach GP how real-world software is built by learning the relevant properties of mined human-coded software - which can be easily accessed through repository hosting services such as GitHub. So combining program synthesis and repository mining is a logical step. In this paper, we analyze if GP can write programs with properties similar to code produced by human software developers. First, we compare the structure of functions generated by different GP initialization methods to a mined corpus containing real-world software. The results show that the studied GP initialization methods produce a totally different combination of programming language elements in comparison to real-world software. Second, we propose perplexity pressure and analyze how its use changes the properties of code produced by GP. The results are very promising and show that we can guide the search to the desired program structure. Thus, we recommend using perplexity pressure as it can be easily integrated in various search-based algorithms.
Autoren
Autoren-URL
https://dl.acm.org/doi/10.1145/3321707.3321738?cid=81100197550
DOI
10.1145/3321707.3321738?cid=81100197550
Conference finish date
2019
ISBN-13
9781450361118
Zeitschrift
GECCO '19: Proceedings of the Genetic and Evolutionary Computation Conference
Conference place
Prague
Name of conference
Genetic and Evolutionary Computation Conference 2019
Paginierung
1065 - 1074, 10
Ort der Veröffentlichung
https://dl.acm.org/doi/10.1145/3321707.3321738?cid=81100197550
Datum der Veröffentlichung
2019
Status
Published
Herausgeber
ACM
Herausgeber URL
https://dl.acm.org/doi/10.1145/3321707.3321738?cid=81100197550
Datum der Datenerfassung
2020
Conference start date
2019
Titel
Teaching GP to program like a human software developer: Using Perplexity Pressure to Guide Program Synthesis Approaches

Data source: Manual

Beziehungen:
Property of

Tools