Epistemic and aleatoric uncertainty in modeling

Daniel J. Segalman, Matthew R. Brake, Lawrence A. Bergman, Alexander F. Vakakis, Kai Willner

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

Abstract

One major difficulty that exists in reconciling model predic-tions of a system with experimental measurements is assessing and accounting for the uncertainties in the system. There are several enumerated sources of uncertainty in model prediction of physical phenomena, the primary ones being: 1) Model form er-ror, 2) Aleatoric uncertainty of model parameters, 3) Epistemic uncertainty of model parameters, and 4) Model solution error. These forms of uncertainty can have insidious consequences for modeling if not properly identified and accounted for. In partic-ular, confusion between aleatoric and epistemic uncertainty can lead to a fundamentally incorrect model being inappropriately fit to data such that the model seems to be correct. As a con-sequence, model predictions may be nonphysical or nonsensicaloutside of the regime for which the model was calibrated. This re-search looks at the effects of aleatoric and epistemic uncertainty in order to make recommendations for properly accounting for them in a modeling framework.

Original languageEnglish (US)
Title of host publication22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise
PublisherAmerican Society of Mechanical Engineers
ISBN (Print)9780791855997
DOIs
StatePublished - Jan 1 2013
EventASME 2013 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2013 - Portland, OR, United States
Duration: Aug 4 2013Aug 7 2013

Publication series

NameProceedings of the ASME Design Engineering Technical Conference
Volume8

Other

OtherASME 2013 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2013
CountryUnited States
CityPortland, OR
Period8/4/138/7/13

Fingerprint

Uncertainty
Epistemic Uncertainty
Modeling
Model
Prediction Model
Recommendations
Form

ASJC Scopus subject areas

  • Modeling and Simulation
  • Mechanical Engineering
  • Computer Science Applications
  • Computer Graphics and Computer-Aided Design

Cite this

Segalman, D. J., Brake, M. R., Bergman, L. A., Vakakis, A. F., & Willner, K. (2013). Epistemic and aleatoric uncertainty in modeling. In 22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise (Proceedings of the ASME Design Engineering Technical Conference; Vol. 8). American Society of Mechanical Engineers. https://doi.org/10.1115/DETC2013-13234

Epistemic and aleatoric uncertainty in modeling. / Segalman, Daniel J.; Brake, Matthew R.; Bergman, Lawrence A.; Vakakis, Alexander F.; Willner, Kai.

22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise. American Society of Mechanical Engineers, 2013. (Proceedings of the ASME Design Engineering Technical Conference; Vol. 8).

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

Segalman, DJ, Brake, MR, Bergman, LA, Vakakis, AF & Willner, K 2013, Epistemic and aleatoric uncertainty in modeling. in 22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise. Proceedings of the ASME Design Engineering Technical Conference, vol. 8, American Society of Mechanical Engineers, ASME 2013 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, IDETC/CIE 2013, Portland, OR, United States, 8/4/13. https://doi.org/10.1115/DETC2013-13234
Segalman DJ, Brake MR, Bergman LA, Vakakis AF, Willner K. Epistemic and aleatoric uncertainty in modeling. In 22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise. American Society of Mechanical Engineers. 2013. (Proceedings of the ASME Design Engineering Technical Conference). https://doi.org/10.1115/DETC2013-13234
Segalman, Daniel J. ; Brake, Matthew R. ; Bergman, Lawrence A. ; Vakakis, Alexander F. ; Willner, Kai. / Epistemic and aleatoric uncertainty in modeling. 22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise. American Society of Mechanical Engineers, 2013. (Proceedings of the ASME Design Engineering Technical Conference).
@inproceedings{5b0755ac815f4aea9e28704c8f7c3b85,
title = "Epistemic and aleatoric uncertainty in modeling",
abstract = "One major difficulty that exists in reconciling model predic-tions of a system with experimental measurements is assessing and accounting for the uncertainties in the system. There are several enumerated sources of uncertainty in model prediction of physical phenomena, the primary ones being: 1) Model form er-ror, 2) Aleatoric uncertainty of model parameters, 3) Epistemic uncertainty of model parameters, and 4) Model solution error. These forms of uncertainty can have insidious consequences for modeling if not properly identified and accounted for. In partic-ular, confusion between aleatoric and epistemic uncertainty can lead to a fundamentally incorrect model being inappropriately fit to data such that the model seems to be correct. As a con-sequence, model predictions may be nonphysical or nonsensicaloutside of the regime for which the model was calibrated. This re-search looks at the effects of aleatoric and epistemic uncertainty in order to make recommendations for properly accounting for them in a modeling framework.",
author = "Segalman, {Daniel J.} and Brake, {Matthew R.} and Bergman, {Lawrence A.} and Vakakis, {Alexander F.} and Kai Willner",
year = "2013",
month = "1",
day = "1",
doi = "10.1115/DETC2013-13234",
language = "English (US)",
isbn = "9780791855997",
series = "Proceedings of the ASME Design Engineering Technical Conference",
publisher = "American Society of Mechanical Engineers",
booktitle = "22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise",

}

TY - GEN

T1 - Epistemic and aleatoric uncertainty in modeling

AU - Segalman, Daniel J.

AU - Brake, Matthew R.

AU - Bergman, Lawrence A.

AU - Vakakis, Alexander F.

AU - Willner, Kai

PY - 2013/1/1

Y1 - 2013/1/1

N2 - One major difficulty that exists in reconciling model predic-tions of a system with experimental measurements is assessing and accounting for the uncertainties in the system. There are several enumerated sources of uncertainty in model prediction of physical phenomena, the primary ones being: 1) Model form er-ror, 2) Aleatoric uncertainty of model parameters, 3) Epistemic uncertainty of model parameters, and 4) Model solution error. These forms of uncertainty can have insidious consequences for modeling if not properly identified and accounted for. In partic-ular, confusion between aleatoric and epistemic uncertainty can lead to a fundamentally incorrect model being inappropriately fit to data such that the model seems to be correct. As a con-sequence, model predictions may be nonphysical or nonsensicaloutside of the regime for which the model was calibrated. This re-search looks at the effects of aleatoric and epistemic uncertainty in order to make recommendations for properly accounting for them in a modeling framework.

AB - One major difficulty that exists in reconciling model predic-tions of a system with experimental measurements is assessing and accounting for the uncertainties in the system. There are several enumerated sources of uncertainty in model prediction of physical phenomena, the primary ones being: 1) Model form er-ror, 2) Aleatoric uncertainty of model parameters, 3) Epistemic uncertainty of model parameters, and 4) Model solution error. These forms of uncertainty can have insidious consequences for modeling if not properly identified and accounted for. In partic-ular, confusion between aleatoric and epistemic uncertainty can lead to a fundamentally incorrect model being inappropriately fit to data such that the model seems to be correct. As a con-sequence, model predictions may be nonphysical or nonsensicaloutside of the regime for which the model was calibrated. This re-search looks at the effects of aleatoric and epistemic uncertainty in order to make recommendations for properly accounting for them in a modeling framework.

UR - http://www.scopus.com/inward/record.url?scp=84896987677&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896987677&partnerID=8YFLogxK

U2 - 10.1115/DETC2013-13234

DO - 10.1115/DETC2013-13234

M3 - Conference contribution

AN - SCOPUS:84896987677

SN - 9780791855997

T3 - Proceedings of the ASME Design Engineering Technical Conference

BT - 22nd Reliability, Stress Analysis, and Failure Prevention Conference; 25th Conference on Mechanical Vibration and Noise

PB - American Society of Mechanical Engineers

ER -