PSense: Automatic Sensitivity Analysis for Probabilistic Programs

Zixin Huang, Zhenbang Wang, Sasa Misailovic

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

PSense is a novel system for sensitivity analysis of probabilistic programs. It computes the impact that a noise in the values of the parameters of the prior distributions and the data have on the program’s result. PSense relates the program executions with and without noise using a developer-provided sensitivity metric. PSense calculates the impact as a set of symbolic functions of each noise variable and supports various non-linear sensitivity metrics. Our evaluation on 66 programs from the literature and five common sensitivity metrics demonstrates the effectiveness of PSense.

Original languageEnglish (US)
Title of host publicationAutomated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings
EditorsChao Wang, Shuvendu K. Lahiri
PublisherSpringer-Verlag
Pages387-403
Number of pages17
ISBN (Print)9783030010898
DOIs
StatePublished - Jan 1 2018
Event16th International Symposium on Automated Technology for Verification and Analysis, ATVA 2018 - Los Angeles, United States
Duration: Oct 7 2018Oct 10 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11138 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other16th International Symposium on Automated Technology for Verification and Analysis, ATVA 2018
CountryUnited States
CityLos Angeles
Period10/7/1810/10/18

Fingerprint

Sensitivity analysis
Sensitivity Analysis
Metric
Prior distribution
Calculate
Evaluation
Demonstrate

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Huang, Z., Wang, Z., & Misailovic, S. (2018). PSense: Automatic Sensitivity Analysis for Probabilistic Programs. In C. Wang, & S. K. Lahiri (Eds.), Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings (pp. 387-403). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11138 LNCS). Springer-Verlag. https://doi.org/10.1007/978-3-030-01090-4_23

PSense : Automatic Sensitivity Analysis for Probabilistic Programs. / Huang, Zixin; Wang, Zhenbang; Misailovic, Sasa.

Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings. ed. / Chao Wang; Shuvendu K. Lahiri. Springer-Verlag, 2018. p. 387-403 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11138 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Huang, Z, Wang, Z & Misailovic, S 2018, PSense: Automatic Sensitivity Analysis for Probabilistic Programs. in C Wang & SK Lahiri (eds), Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11138 LNCS, Springer-Verlag, pp. 387-403, 16th International Symposium on Automated Technology for Verification and Analysis, ATVA 2018, Los Angeles, United States, 10/7/18. https://doi.org/10.1007/978-3-030-01090-4_23
Huang Z, Wang Z, Misailovic S. PSense: Automatic Sensitivity Analysis for Probabilistic Programs. In Wang C, Lahiri SK, editors, Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings. Springer-Verlag. 2018. p. 387-403. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-01090-4_23
Huang, Zixin ; Wang, Zhenbang ; Misailovic, Sasa. / PSense : Automatic Sensitivity Analysis for Probabilistic Programs. Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings. editor / Chao Wang ; Shuvendu K. Lahiri. Springer-Verlag, 2018. pp. 387-403 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2fdaf180b2864076b1c0dbeeb76e7b92,
title = "PSense: Automatic Sensitivity Analysis for Probabilistic Programs",
abstract = "PSense is a novel system for sensitivity analysis of probabilistic programs. It computes the impact that a noise in the values of the parameters of the prior distributions and the data have on the program’s result. PSense relates the program executions with and without noise using a developer-provided sensitivity metric. PSense calculates the impact as a set of symbolic functions of each noise variable and supports various non-linear sensitivity metrics. Our evaluation on 66 programs from the literature and five common sensitivity metrics demonstrates the effectiveness of PSense.",
author = "Zixin Huang and Zhenbang Wang and Sasa Misailovic",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-030-01090-4_23",
language = "English (US)",
isbn = "9783030010898",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer-Verlag",
pages = "387--403",
editor = "Chao Wang and Lahiri, {Shuvendu K.}",
booktitle = "Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings",

}

TY - GEN

T1 - PSense

T2 - Automatic Sensitivity Analysis for Probabilistic Programs

AU - Huang, Zixin

AU - Wang, Zhenbang

AU - Misailovic, Sasa

PY - 2018/1/1

Y1 - 2018/1/1

N2 - PSense is a novel system for sensitivity analysis of probabilistic programs. It computes the impact that a noise in the values of the parameters of the prior distributions and the data have on the program’s result. PSense relates the program executions with and without noise using a developer-provided sensitivity metric. PSense calculates the impact as a set of symbolic functions of each noise variable and supports various non-linear sensitivity metrics. Our evaluation on 66 programs from the literature and five common sensitivity metrics demonstrates the effectiveness of PSense.

AB - PSense is a novel system for sensitivity analysis of probabilistic programs. It computes the impact that a noise in the values of the parameters of the prior distributions and the data have on the program’s result. PSense relates the program executions with and without noise using a developer-provided sensitivity metric. PSense calculates the impact as a set of symbolic functions of each noise variable and supports various non-linear sensitivity metrics. Our evaluation on 66 programs from the literature and five common sensitivity metrics demonstrates the effectiveness of PSense.

UR - http://www.scopus.com/inward/record.url?scp=85054879346&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054879346&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-01090-4_23

DO - 10.1007/978-3-030-01090-4_23

M3 - Conference contribution

AN - SCOPUS:85054879346

SN - 9783030010898

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 387

EP - 403

BT - Automated Technology for Verification and Analysis - 16th International Symposium, ATVA 2018, Proceedings

A2 - Wang, Chao

A2 - Lahiri, Shuvendu K.

PB - Springer-Verlag

ER -