Robust multitask Elliptical Regression (ROMER)

Nofar Noy; Yaav Wald; Gal Elidan; Ami Wiesel

doi:10.1109/CAMSAP45676.2019.9022524

Robust multitask Elliptical Regression (ROMER)

Nofar Noy, Yaav Wald, Gal Elidan, Ami Wiesel

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

1 Scopus citations

Abstract

Multitask regression, and in particular the setting where each task is data starved, is a common challenge in machine learning and signal processing. The goal in such cases is to leverage the correlation or joint structure between the tasks in order to gain higher accuracy. In realistic settings the underlying distribution is often heavy-tailed or contaminated and requires the use of robust statistics. To enjoy the benefits of both, we interpret multitask regression as estimation of the parameters of a multivariate conditional distribution. This naturally leads to RObust Multitask Elliptical Regression (ROMER), and allows for robust estimation in the predictive or conditional multitask setting. Following the introduction of the model, we characterize the optimization landscape, and demonstrate its efficacy in a real-world problem of river discharge estimation across multiple river sites.

Original language	English
Title of host publication	2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	261-265
Number of pages	5
ISBN (Electronic)	9781728155494
DOIs	https://doi.org/10.1109/CAMSAP45676.2019.9022524
State	Published - Dec 2019
Event	8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Le Gosier, Guadeloupe Duration: 15 Dec 2019 → 18 Dec 2019

Publication series

Name	2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings

Conference

Conference	8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019
Country/Territory	Guadeloupe
City	Le Gosier
Period	15/12/19 → 18/12/19

Bibliographical note

Publisher Copyright:
© 2019 IEEE.

Keywords

Robust regression
conditional graphical models
geodesic convexity
multivariate elliptical distributions

Access to Document

10.1109/CAMSAP45676.2019.9022524

Cite this

Noy, N., Wald, Y., Elidan, G., & Wiesel, A. (2019). Robust multitask Elliptical Regression (ROMER). In 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings (pp. 261-265). Article 9022524 (2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CAMSAP45676.2019.9022524

Noy, Nofar ; Wald, Yaav ; Elidan, Gal et al. / Robust multitask Elliptical Regression (ROMER). 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 261-265 (2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings).

@inproceedings{4aee6a0dc19442ea8281e5e999bece15,

title = "Robust multitask Elliptical Regression (ROMER)",

abstract = "Multitask regression, and in particular the setting where each task is data starved, is a common challenge in machine learning and signal processing. The goal in such cases is to leverage the correlation or joint structure between the tasks in order to gain higher accuracy. In realistic settings the underlying distribution is often heavy-tailed or contaminated and requires the use of robust statistics. To enjoy the benefits of both, we interpret multitask regression as estimation of the parameters of a multivariate conditional distribution. This naturally leads to RObust Multitask Elliptical Regression (ROMER), and allows for robust estimation in the predictive or conditional multitask setting. Following the introduction of the model, we characterize the optimization landscape, and demonstrate its efficacy in a real-world problem of river discharge estimation across multiple river sites.",

keywords = "Robust regression, conditional graphical models, geodesic convexity, multivariate elliptical distributions",

author = "Nofar Noy and Yaav Wald and Gal Elidan and Ami Wiesel",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 ; Conference date: 15-12-2019 Through 18-12-2019",

year = "2019",

month = dec,

doi = "10.1109/CAMSAP45676.2019.9022524",

language = "אנגלית",

series = "2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "261--265",

booktitle = "2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings",

address = "ארצות הברית",

}

Noy, N, Wald, Y, Elidan, G & Wiesel, A 2019, Robust multitask Elliptical Regression (ROMER). in 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings., 9022524, 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 261-265, 8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019, Le Gosier, Guadeloupe, 15/12/19. https://doi.org/10.1109/CAMSAP45676.2019.9022524

Robust multitask Elliptical Regression (ROMER). / Noy, Nofar; Wald, Yaav; Elidan, Gal et al.
2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 261-265 9022524 (2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings).

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

TY - GEN

T1 - Robust multitask Elliptical Regression (ROMER)

AU - Noy, Nofar

AU - Wald, Yaav

AU - Elidan, Gal

AU - Wiesel, Ami

PY - 2019/12

Y1 - 2019/12

N2 - Multitask regression, and in particular the setting where each task is data starved, is a common challenge in machine learning and signal processing. The goal in such cases is to leverage the correlation or joint structure between the tasks in order to gain higher accuracy. In realistic settings the underlying distribution is often heavy-tailed or contaminated and requires the use of robust statistics. To enjoy the benefits of both, we interpret multitask regression as estimation of the parameters of a multivariate conditional distribution. This naturally leads to RObust Multitask Elliptical Regression (ROMER), and allows for robust estimation in the predictive or conditional multitask setting. Following the introduction of the model, we characterize the optimization landscape, and demonstrate its efficacy in a real-world problem of river discharge estimation across multiple river sites.

AB - Multitask regression, and in particular the setting where each task is data starved, is a common challenge in machine learning and signal processing. The goal in such cases is to leverage the correlation or joint structure between the tasks in order to gain higher accuracy. In realistic settings the underlying distribution is often heavy-tailed or contaminated and requires the use of robust statistics. To enjoy the benefits of both, we interpret multitask regression as estimation of the parameters of a multivariate conditional distribution. This naturally leads to RObust Multitask Elliptical Regression (ROMER), and allows for robust estimation in the predictive or conditional multitask setting. Following the introduction of the model, we characterize the optimization landscape, and demonstrate its efficacy in a real-world problem of river discharge estimation across multiple river sites.

KW - Robust regression

KW - conditional graphical models

KW - geodesic convexity

KW - multivariate elliptical distributions

UR - http://www.scopus.com/inward/record.url?scp=85082391126&partnerID=8YFLogxK

U2 - 10.1109/CAMSAP45676.2019.9022524

DO - 10.1109/CAMSAP45676.2019.9022524

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85082391126

T3 - 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings

SP - 261

EP - 265

BT - 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 8th IEEE International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019

Y2 - 15 December 2019 through 18 December 2019

ER -

Noy N, Wald Y, Elidan G , Wiesel A. Robust multitask Elliptical Regression (ROMER). In 2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 261-265. 9022524. (2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing, CAMSAP 2019 - Proceedings). doi: 10.1109/CAMSAP45676.2019.9022524

Robust multitask Elliptical Regression (ROMER)

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this