Improving model inference in industry by combining active and passive learning

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

1 Citation (Scopus)

Abstract

Inferring behavioral models (e.g., state machines) of software systems is an important element of re-engineering activities. Model inference techniques can be categorized as active or passive learning, constructing models by (dynamically) interacting with systems or (statically) analyzing traces, respectively. Application of those techniques in the industry is, however, hindered by the trade-off between learning time and completeness achieved (active learning) or by incomplete input logs (passive learning). We investigate the learning time/completeness achieved trade-off of active learning with a pilot study at ASML, provider of lithography systems for the semiconductor industry. To resolve the trade-off we advocate extending active learning with execution logs and passive learning results.We apply the extended approach to eighteen components used in ASML TWINSCAN lithography machines. Compared to traditional active learning, our approach significantly reduces the active learning time. Moreover, it is capable of learning the behavior missed by the traditional active learning approach.

Original languageEnglish
Title of host publication26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019)
EditorsEmad Shihab, David Lo, Xinyu Wang
Place of PublicationPiscataway
PublisherInstitute of Electrical and Electronics Engineers
Pages253-263
Number of pages11
ISBN (Electronic)9781728105918
DOIs
Publication statusPublished - 15 Mar 2019
Event26th IEEE International Conference on Software Analysis, Evolution, and Reengineering, SANER 2019 - Hangzhou, China
Duration: 24 Feb 201927 Feb 2019
Conference number: 26

Conference

Conference26th IEEE International Conference on Software Analysis, Evolution, and Reengineering, SANER 2019
Abbreviated titleSANER
CountryChina
CityHangzhou
Period24/02/1927/02/19

Fingerprint

Industry
Lithography
Problem-Based Learning
Semiconductor materials

Keywords

  • active learning
  • equivalence oracle
  • model inference
  • passive learning
  • reverse engineering
  • runtime logs

Cite this

Yang, N., Aslam, K., Schiffelers, R. R. H., Lensink, L., Hendriks, D., Cleophas, L. G. W. A., & Serebrenik, A. (2019). Improving model inference in industry by combining active and passive learning. In E. Shihab, D. Lo, & X. Wang (Eds.), 26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019) (pp. 253-263). [8668007] Piscataway: Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/SANER.2019.8668007
Yang, Nan ; Aslam, K. ; Schiffelers, R.R.H. ; Lensink, Leonard ; Hendriks, D. ; Cleophas, L.G.W.A. ; Serebrenik, A. / Improving model inference in industry by combining active and passive learning. 26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019). editor / Emad Shihab ; David Lo ; Xinyu Wang. Piscataway : Institute of Electrical and Electronics Engineers, 2019. pp. 253-263
@inproceedings{2bb9e63a37514d0d9b505f34d2bfe106,
title = "Improving model inference in industry by combining active and passive learning",
abstract = "Inferring behavioral models (e.g., state machines) of software systems is an important element of re-engineering activities. Model inference techniques can be categorized as active or passive learning, constructing models by (dynamically) interacting with systems or (statically) analyzing traces, respectively. Application of those techniques in the industry is, however, hindered by the trade-off between learning time and completeness achieved (active learning) or by incomplete input logs (passive learning). We investigate the learning time/completeness achieved trade-off of active learning with a pilot study at ASML, provider of lithography systems for the semiconductor industry. To resolve the trade-off we advocate extending active learning with execution logs and passive learning results.We apply the extended approach to eighteen components used in ASML TWINSCAN lithography machines. Compared to traditional active learning, our approach significantly reduces the active learning time. Moreover, it is capable of learning the behavior missed by the traditional active learning approach.",
keywords = "active learning, equivalence oracle, model inference, passive learning, reverse engineering, runtime logs",
author = "Nan Yang and K. Aslam and R.R.H. Schiffelers and Leonard Lensink and D. Hendriks and L.G.W.A. Cleophas and A. Serebrenik",
year = "2019",
month = "3",
day = "15",
doi = "10.1109/SANER.2019.8668007",
language = "English",
pages = "253--263",
editor = "Emad Shihab and David Lo and Xinyu Wang",
booktitle = "26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019)",
publisher = "Institute of Electrical and Electronics Engineers",
address = "United States",

}

Yang, N, Aslam, K, Schiffelers, RRH, Lensink, L, Hendriks, D, Cleophas, LGWA & Serebrenik, A 2019, Improving model inference in industry by combining active and passive learning. in E Shihab, D Lo & X Wang (eds), 26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019)., 8668007, Institute of Electrical and Electronics Engineers, Piscataway, pp. 253-263, 26th IEEE International Conference on Software Analysis, Evolution, and Reengineering, SANER 2019, Hangzhou, China, 24/02/19. https://doi.org/10.1109/SANER.2019.8668007

Improving model inference in industry by combining active and passive learning. / Yang, Nan; Aslam, K.; Schiffelers, R.R.H.; Lensink, Leonard; Hendriks, D.; Cleophas, L.G.W.A.; Serebrenik, A.

26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019). ed. / Emad Shihab; David Lo; Xinyu Wang. Piscataway : Institute of Electrical and Electronics Engineers, 2019. p. 253-263 8668007.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Improving model inference in industry by combining active and passive learning

AU - Yang, Nan

AU - Aslam, K.

AU - Schiffelers, R.R.H.

AU - Lensink, Leonard

AU - Hendriks, D.

AU - Cleophas, L.G.W.A.

AU - Serebrenik, A.

PY - 2019/3/15

Y1 - 2019/3/15

N2 - Inferring behavioral models (e.g., state machines) of software systems is an important element of re-engineering activities. Model inference techniques can be categorized as active or passive learning, constructing models by (dynamically) interacting with systems or (statically) analyzing traces, respectively. Application of those techniques in the industry is, however, hindered by the trade-off between learning time and completeness achieved (active learning) or by incomplete input logs (passive learning). We investigate the learning time/completeness achieved trade-off of active learning with a pilot study at ASML, provider of lithography systems for the semiconductor industry. To resolve the trade-off we advocate extending active learning with execution logs and passive learning results.We apply the extended approach to eighteen components used in ASML TWINSCAN lithography machines. Compared to traditional active learning, our approach significantly reduces the active learning time. Moreover, it is capable of learning the behavior missed by the traditional active learning approach.

AB - Inferring behavioral models (e.g., state machines) of software systems is an important element of re-engineering activities. Model inference techniques can be categorized as active or passive learning, constructing models by (dynamically) interacting with systems or (statically) analyzing traces, respectively. Application of those techniques in the industry is, however, hindered by the trade-off between learning time and completeness achieved (active learning) or by incomplete input logs (passive learning). We investigate the learning time/completeness achieved trade-off of active learning with a pilot study at ASML, provider of lithography systems for the semiconductor industry. To resolve the trade-off we advocate extending active learning with execution logs and passive learning results.We apply the extended approach to eighteen components used in ASML TWINSCAN lithography machines. Compared to traditional active learning, our approach significantly reduces the active learning time. Moreover, it is capable of learning the behavior missed by the traditional active learning approach.

KW - active learning

KW - equivalence oracle

KW - model inference

KW - passive learning

KW - reverse engineering

KW - runtime logs

UR - http://www.scopus.com/inward/record.url?scp=85064164636&partnerID=8YFLogxK

U2 - 10.1109/SANER.2019.8668007

DO - 10.1109/SANER.2019.8668007

M3 - Conference contribution

AN - SCOPUS:85064164636

SP - 253

EP - 263

BT - 26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019)

A2 - Shihab, Emad

A2 - Lo, David

A2 - Wang, Xinyu

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

ER -

Yang N, Aslam K, Schiffelers RRH, Lensink L, Hendriks D, Cleophas LGWA et al. Improving model inference in industry by combining active and passive learning. In Shihab E, Lo D, Wang X, editors, 26th IEEE International Conference on Software Analysis, Evolution, and Reengineering (SANER 2019). Piscataway: Institute of Electrical and Electronics Engineers. 2019. p. 253-263. 8668007 https://doi.org/10.1109/SANER.2019.8668007