DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search

Alessio Gambi; Paolo Tonella; Vincenzo Riccio; Tahereh Zohdinasab

DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search

Alessio Gambi
, Paolo Tonella
, Vincenzo Riccio
, Tahereh Zohdinasab

Institut Digitalisierung und Informatik

Università della Svizzera italiana
Software Institute - USI

Publikation: Beitrag in Buch/Bericht/Konferenzband › Konferenzbeitrag › Begutachtung

Abstract

In this extended abstract, we summarize our contributions to automated testing of Deep Learning-based systems published at the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) in 2021 [Zo21a] and just accepted by the ACM Transactions on Software Engineering and Methodology (TOSEM) in 2022 [Zo22]. Deep Learning-based systems (DL Systems) find applications in safety-critical application domains and thus must be thoroughly tested. Existing DL system testing approaches can generate complex and fault-finding inputs but do not characterize them in a way that enables human interpretation and do not always consider test diversity. Our work addresses these challenges and can find effective and diverse test cases.

Originalsprache	Englisch
Titel	Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik
Redakteure/-innen	Gregor Engels, Regina Hebig, Matthias Tichy
Herausgeber (Verlag)	Gesellschaft für Informatik e.V.
Seiten	131-132
Seitenumfang	2
Band	P-332
ISBN (elektronisch)	9783885797265
Publikationsstatus	Veröffentlicht - 24 Feb. 2023

Publikationsreihe

Name	Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)
Band	P-332
ISSN (Print)	1617-5468

UN SDGs

Dieser Output leistet einen Beitrag zu folgendem(n) Ziel(en) für nachhaltige Entwicklung

SDG 11 – Nachhaltige Städte und Gemeinschaften

IMC Forschungsschwerpunkte

Software engineering and intelligent systems

ÖFOS 2012 - Österreichischen Systematik der Wissenschaftszweige

102015 Informationssysteme

Andere Dateien und Links

Verknüpfung zur Publikation in Scopus

Dieses zitieren

Gambi, A., Tonella, P., Riccio, V., & Zohdinasab, T. (2023). DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search. in G. Engels, R. Hebig, & M. Tichy (Hrsg.), Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik (Band P-332, S. 131-132). (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI); Band P-332). Gesellschaft für Informatik e.V..

Gambi, Alessio ; Tonella, Paolo ; Riccio, Vincenzo et al. / DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search. Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik. Hrsg. / Gregor Engels ; Regina Hebig ; Matthias Tichy. Band P-332 Gesellschaft für Informatik e.V., 2023. S. 131-132 (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)).

@inproceedings{fe69fa7866664613bcd7bc1715e16a97,

title = "DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search",

abstract = "In this extended abstract, we summarize our contributions to automated testing of Deep Learning-based systems published at the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) in 2021 [Zo21a] and just accepted by the ACM Transactions on Software Engineering and Methodology (TOSEM) in 2022 [Zo22]. Deep Learning-based systems (DL Systems) find applications in safety-critical application domains and thus must be thoroughly tested. Existing DL system testing approaches can generate complex and fault-finding inputs but do not characterize them in a way that enables human interpretation and do not always consider test diversity. Our work addresses these challenges and can find effective and diverse test cases.",

keywords = "deep learning, search-based software engineering, self-driving cars, Software testing",

author = "Alessio Gambi and Paolo Tonella and Vincenzo Riccio and Tahereh Zohdinasab",

year = "2023",

month = feb,

day = "24",

language = "English",

volume = "P-332",

series = "Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)",

publisher = "Gesellschaft f{\"u}r Informatik e.V.",

pages = "131--132",

editor = "Gregor Engels and Regina Hebig and Matthias Tichy",

booktitle = "Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik",

}

Gambi, A, Tonella, P, Riccio, V & Zohdinasab, T 2023, DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search. in G Engels, R Hebig & M Tichy (Hrsg.), Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik. Bd. P-332, Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI), Bd. P-332, Gesellschaft für Informatik e.V., S. 131-132.

DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search. / Gambi, Alessio; Tonella, Paolo; Riccio, Vincenzo et al.
Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik. Hrsg. / Gregor Engels; Regina Hebig; Matthias Tichy. Band P-332 Gesellschaft für Informatik e.V., 2023. S. 131-132 (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI); Band P-332).

Publikation: Beitrag in Buch/Bericht/Konferenzband › Konferenzbeitrag › Begutachtung

TY - GEN

T1 - DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search

AU - Gambi, Alessio

AU - Tonella, Paolo

AU - Riccio, Vincenzo

AU - Zohdinasab, Tahereh

PY - 2023/2/24

Y1 - 2023/2/24

N2 - In this extended abstract, we summarize our contributions to automated testing of Deep Learning-based systems published at the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) in 2021 [Zo21a] and just accepted by the ACM Transactions on Software Engineering and Methodology (TOSEM) in 2022 [Zo22]. Deep Learning-based systems (DL Systems) find applications in safety-critical application domains and thus must be thoroughly tested. Existing DL system testing approaches can generate complex and fault-finding inputs but do not characterize them in a way that enables human interpretation and do not always consider test diversity. Our work addresses these challenges and can find effective and diverse test cases.

AB - In this extended abstract, we summarize our contributions to automated testing of Deep Learning-based systems published at the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) in 2021 [Zo21a] and just accepted by the ACM Transactions on Software Engineering and Methodology (TOSEM) in 2022 [Zo22]. Deep Learning-based systems (DL Systems) find applications in safety-critical application domains and thus must be thoroughly tested. Existing DL system testing approaches can generate complex and fault-finding inputs but do not characterize them in a way that enables human interpretation and do not always consider test diversity. Our work addresses these challenges and can find effective and diverse test cases.

KW - deep learning

KW - search-based software engineering

KW - self-driving cars

KW - Software testing

UR - https://www.scopus.com/pages/publications/85150057352

M3 - Conference contribution

VL - P-332

T3 - Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)

SP - 131

EP - 132

BT - Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik

A2 - Engels, Gregor

A2 - Hebig, Regina

A2 - Tichy, Matthias

PB - Gesellschaft für Informatik e.V.

ER -

Gambi A, Tonella P, Riccio V, Zohdinasab T. DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search. in Engels G, Hebig R, Tichy M, Hrsg., Software Engineering 2023 - Fachtagung des GI-Fachbereichs Softwaretechnik. Band P-332. Gesellschaft für Informatik e.V. 2023. S. 131-132. (Lecture Notes in Informatics (LNI), Proceedings - Series of the Gesellschaft fur Informatik (GI)).

DeepHyperion: Exploring the Feature Space of Deep Learning-based Systems through Illumination Search

Abstract

Publikationsreihe

UN SDGs

IMC Forschungsschwerpunkte

ÖFOS 2012 - Österreichischen Systematik der Wissenschaftszweige

Andere Dateien und Links

Fingerprint

Dieses zitieren