Bayesian Crowdsourcing with Constraints

Panagiotis A. Traganitis; Georgios B. Giannakis

doi:10.1007/978-3-030-86523-8_33

Bayesian Crowdsourcing with Constraints

Panagiotis A. Traganitis, Georgios B. Giannakis

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Crowdsourcing has emerged as a powerful paradigm for efficiently labeling large datasets and performing various learning tasks, by leveraging crowds of human annotators. When additional information is available about the data, constrained or semi-supervised crowdsourcing approaches that enhance the aggregation of labels from human annotators are well motivated. This work deals with constrained crowdsourced classification with instance-level constraints, that capture relationships between pairs of data. A Bayesian algorithm based on variational inference is developed, and its quantifiably improved performance, compared to unsupervised crowdsourcing, is analytically and empirically validated on several crowdsourcing datasets.

Original language	English (US)
Title of host publication	Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings
Editors	Nuria Oliver, Fernando Pérez-Cruz, Stefan Kramer, Jesse Read, Jose A. Lozano
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	543-559
Number of pages	17
ISBN (Print)	9783030865221
DOIs	https://doi.org/10.1007/978-3-030-86523-8_33
State	Published - 2021
Event	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021 - Virtual, Online Duration: Sep 13 2021 → Sep 17 2021

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12977 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021
City	Virtual, Online
Period	9/13/21 → 9/17/21

Bibliographical note

Publisher Copyright:
© 2021, Springer Nature Switzerland AG.

Keywords

Bayesian
Crowdsourcing
Ensemble learning
Semi-supervised
Variational inference

Publisher link

10.1007/978-3-030-86523-8_33

Find It

Find It at U of M Libraries

Cite this

Traganitis, P. A., & Giannakis, G. B. (2021). Bayesian Crowdsourcing with Constraints. In N. Oliver, F. Pérez-Cruz, S. Kramer, J. Read, & J. A. Lozano (Eds.), Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings (pp. 543-559). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12977 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-86523-8_33

Bayesian Crowdsourcing with Constraints. / Traganitis, Panagiotis A.; Giannakis, Georgios B.
Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings. ed. / Nuria Oliver; Fernando Pérez-Cruz; Stefan Kramer; Jesse Read; Jose A. Lozano. Springer Science and Business Media Deutschland GmbH, 2021. p. 543-559 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12977 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Traganitis, PA & Giannakis, GB 2021, Bayesian Crowdsourcing with Constraints. in N Oliver, F Pérez-Cruz, S Kramer, J Read & JA Lozano (eds), Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12977 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 543-559, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021, Virtual, Online, 9/13/21. https://doi.org/10.1007/978-3-030-86523-8_33

Traganitis PA, Giannakis GB. Bayesian Crowdsourcing with Constraints. In Oliver N, Pérez-Cruz F, Kramer S, Read J, Lozano JA, editors, Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 543-559. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-86523-8_33

Traganitis, Panagiotis A. ; Giannakis, Georgios B. / Bayesian Crowdsourcing with Constraints. Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings. editor / Nuria Oliver ; Fernando Pérez-Cruz ; Stefan Kramer ; Jesse Read ; Jose A. Lozano. Springer Science and Business Media Deutschland GmbH, 2021. pp. 543-559 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{28810439adaa4933b938ceb34548af77,

title = "Bayesian Crowdsourcing with Constraints",

abstract = "Crowdsourcing has emerged as a powerful paradigm for efficiently labeling large datasets and performing various learning tasks, by leveraging crowds of human annotators. When additional information is available about the data, constrained or semi-supervised crowdsourcing approaches that enhance the aggregation of labels from human annotators are well motivated. This work deals with constrained crowdsourced classification with instance-level constraints, that capture relationships between pairs of data. A Bayesian algorithm based on variational inference is developed, and its quantifiably improved performance, compared to unsupervised crowdsourcing, is analytically and empirically validated on several crowdsourcing datasets.",

keywords = "Bayesian, Crowdsourcing, Ensemble learning, Semi-supervised, Variational inference",

author = "Traganitis, {Panagiotis A.} and Giannakis, {Georgios B.}",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021 ; Conference date: 13-09-2021 Through 17-09-2021",

year = "2021",

doi = "10.1007/978-3-030-86523-8_33",

language = "English (US)",

isbn = "9783030865221",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "543--559",

editor = "Nuria Oliver and Fernando P{\'e}rez-Cruz and Stefan Kramer and Jesse Read and Lozano, {Jose A.}",

booktitle = "Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Bayesian Crowdsourcing with Constraints

AU - Traganitis, Panagiotis A.

AU - Giannakis, Georgios B.

PY - 2021

Y1 - 2021

N2 - Crowdsourcing has emerged as a powerful paradigm for efficiently labeling large datasets and performing various learning tasks, by leveraging crowds of human annotators. When additional information is available about the data, constrained or semi-supervised crowdsourcing approaches that enhance the aggregation of labels from human annotators are well motivated. This work deals with constrained crowdsourced classification with instance-level constraints, that capture relationships between pairs of data. A Bayesian algorithm based on variational inference is developed, and its quantifiably improved performance, compared to unsupervised crowdsourcing, is analytically and empirically validated on several crowdsourcing datasets.

AB - Crowdsourcing has emerged as a powerful paradigm for efficiently labeling large datasets and performing various learning tasks, by leveraging crowds of human annotators. When additional information is available about the data, constrained or semi-supervised crowdsourcing approaches that enhance the aggregation of labels from human annotators are well motivated. This work deals with constrained crowdsourced classification with instance-level constraints, that capture relationships between pairs of data. A Bayesian algorithm based on variational inference is developed, and its quantifiably improved performance, compared to unsupervised crowdsourcing, is analytically and empirically validated on several crowdsourcing datasets.

KW - Bayesian

KW - Crowdsourcing

KW - Ensemble learning

KW - Semi-supervised

KW - Variational inference

UR - http://www.scopus.com/inward/record.url?scp=85115725669&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85115725669&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-86523-8_33

DO - 10.1007/978-3-030-86523-8_33

M3 - Conference contribution

AN - SCOPUS:85115725669

SN - 9783030865221

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

SP - 543

EP - 559

BT - Machine Learning and Knowledge Discovery in Databases. Research Track - European Conference, ECML PKDD 2021, Proceedings

A2 - Oliver, Nuria

A2 - Pérez-Cruz, Fernando

A2 - Kramer, Stefan

A2 - Read, Jesse

A2 - Lozano, Jose A.

PB - Springer Science and Business Media Deutschland GmbH

T2 - European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2021

Y2 - 13 September 2021 through 17 September 2021

ER -

Bayesian Crowdsourcing with Constraints

Abstract

Publication series

Conference

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this