A differential privacy workflow for inference of parameters in the rasch model

Teresa Anna Steiner, David Enslev Nyrnberg, Lars Kai Hansen*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

The Rasch model is used to estimate student performance and task difficulty in simple test scenarios. We design a workflow for enhancing student feedback by release of difficulty parameters in the Rasch model with privacy protection using differential privacy. We provide a first proof of differential privacy in Rasch models and derive the minimum noise level in objective perturbation to guarantee a given privacy budget. We test the workflow in simulations and in two real data sets.

Original languageEnglish
Title of host publicationProceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018
EditorsAnna Monreale, Carlos Alzate
PublisherSpringer
Publication date1 Jan 2019
Pages113-124
ISBN (Print)9783030134624
DOIs
Publication statusPublished - 1 Jan 2019
Event3rd Workshop on Mining Data for Financial Applications, MIDAS 2018 and 2nd International Workshop on Personal Analytics and Privacy, PAP 2018 held at 18th European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2018 - Dublin, Ireland
Duration: 10 Sep 201814 Sep 2018

Conference

Conference3rd Workshop on Mining Data for Financial Applications, MIDAS 2018 and 2nd International Workshop on Personal Analytics and Privacy, PAP 2018 held at 18th European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2018
CountryIreland
CityDublin
Period10/09/201814/09/2018
SeriesLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11054 LNAI
ISSN0302-9743

Keywords

  • Differential privacy
  • Rasch model
  • Student feedback

Cite this

Steiner, T. A., Nyrnberg, D. E., & Hansen, L. K. (2019). A differential privacy workflow for inference of parameters in the rasch model. In A. Monreale, & C. Alzate (Eds.), Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018 (pp. 113-124). Springer. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol.. 11054 LNAI https://doi.org/10.1007/978-3-030-13463-1_9
Steiner, Teresa Anna ; Nyrnberg, David Enslev ; Hansen, Lars Kai. / A differential privacy workflow for inference of parameters in the rasch model. Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018. editor / Anna Monreale ; Carlos Alzate. Springer, 2019. pp. 113-124 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11054 LNAI).
@inproceedings{47f80231d7e24ca6b0e6095cf75309b6,
title = "A differential privacy workflow for inference of parameters in the rasch model",
abstract = "The Rasch model is used to estimate student performance and task difficulty in simple test scenarios. We design a workflow for enhancing student feedback by release of difficulty parameters in the Rasch model with privacy protection using differential privacy. We provide a first proof of differential privacy in Rasch models and derive the minimum noise level in objective perturbation to guarantee a given privacy budget. We test the workflow in simulations and in two real data sets.",
keywords = "Differential privacy, Rasch model, Student feedback",
author = "Steiner, {Teresa Anna} and Nyrnberg, {David Enslev} and Hansen, {Lars Kai}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-13463-1_9",
language = "English",
isbn = "9783030134624",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "113--124",
editor = "Anna Monreale and Carlos Alzate",
booktitle = "Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018",

}

Steiner, TA, Nyrnberg, DE & Hansen, LK 2019, A differential privacy workflow for inference of parameters in the rasch model. in A Monreale & C Alzate (eds), Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018. Springer, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11054 LNAI, pp. 113-124, 3rd Workshop on Mining Data for Financial Applications, MIDAS 2018 and 2nd International Workshop on Personal Analytics and Privacy, PAP 2018 held at 18th European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2018, Dublin, Ireland, 10/09/2018. https://doi.org/10.1007/978-3-030-13463-1_9

A differential privacy workflow for inference of parameters in the rasch model. / Steiner, Teresa Anna; Nyrnberg, David Enslev; Hansen, Lars Kai.

Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018. ed. / Anna Monreale; Carlos Alzate. Springer, 2019. p. 113-124 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11054 LNAI).

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - A differential privacy workflow for inference of parameters in the rasch model

AU - Steiner, Teresa Anna

AU - Nyrnberg, David Enslev

AU - Hansen, Lars Kai

PY - 2019/1/1

Y1 - 2019/1/1

N2 - The Rasch model is used to estimate student performance and task difficulty in simple test scenarios. We design a workflow for enhancing student feedback by release of difficulty parameters in the Rasch model with privacy protection using differential privacy. We provide a first proof of differential privacy in Rasch models and derive the minimum noise level in objective perturbation to guarantee a given privacy budget. We test the workflow in simulations and in two real data sets.

AB - The Rasch model is used to estimate student performance and task difficulty in simple test scenarios. We design a workflow for enhancing student feedback by release of difficulty parameters in the Rasch model with privacy protection using differential privacy. We provide a first proof of differential privacy in Rasch models and derive the minimum noise level in objective perturbation to guarantee a given privacy budget. We test the workflow in simulations and in two real data sets.

KW - Differential privacy

KW - Rasch model

KW - Student feedback

U2 - 10.1007/978-3-030-13463-1_9

DO - 10.1007/978-3-030-13463-1_9

M3 - Article in proceedings

AN - SCOPUS:85063419666

SN - 9783030134624

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

SP - 113

EP - 124

BT - Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018

A2 - Monreale, Anna

A2 - Alzate, Carlos

PB - Springer

ER -

Steiner TA, Nyrnberg DE, Hansen LK. A differential privacy workflow for inference of parameters in the rasch model. In Monreale A, Alzate C, editors, Proceedings of ECML PKDD 2018 Workshops - MIDAS 2018 and PAP 2018. Springer. 2019. p. 113-124. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11054 LNAI). https://doi.org/10.1007/978-3-030-13463-1_9