Online Learning with Abstention

Corinna Cortes, Giulia Desalvo, Claudio Gentile, Mehryar Mohri, Scott Yang

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

Abstract

We present an extensive study of a key problem in online learning where the learner can opt to abstain from making a prediction, at a ccrtain cost. In the adversarial setting, we show how existing online algorithms and guarantees can be adapted to this problem. In the stochastic setting, we first point out a bias problem that limits the straightforward extension of algorithms such as ucb-n to this context. Next, we give a new algorithm, ucb- gt, that exploits historical data and time-varying feedback graphs. We show that this algorithm ben: cfits from more favorable regret guarantees than a natural extension of ucb-n. We further report the results of a series of experiments demonstrating that ucb-gt largely outperforms that extension of ucb-n, as well as other standard baselines.

Original languageEnglish (US)
Title of host publication35th International Conference on Machine Learning, ICML 2018
EditorsAndreas Krause, Jennifer Dy
PublisherInternational Machine Learning Society (IMLS)
Pages1726-1734
Number of pages9
Volume3
ISBN (Electronic)9781510867963
StatePublished - Jan 1 2018
Event35th International Conference on Machine Learning, ICML 2018 - Stockholm, Sweden
Duration: Jul 10 2018Jul 15 2018

Other

Other35th International Conference on Machine Learning, ICML 2018
CountrySweden
CityStockholm
Period7/10/187/15/18

Fingerprint

Feedback
Costs
Experiments

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Human-Computer Interaction
  • Software

Cite this

Cortes, C., Desalvo, G., Gentile, C., Mohri, M., & Yang, S. (2018). Online Learning with Abstention. In A. Krause, & J. Dy (Eds.), 35th International Conference on Machine Learning, ICML 2018 (Vol. 3, pp. 1726-1734). International Machine Learning Society (IMLS).

Online Learning with Abstention. / Cortes, Corinna; Desalvo, Giulia; Gentile, Claudio; Mohri, Mehryar; Yang, Scott.

35th International Conference on Machine Learning, ICML 2018. ed. / Andreas Krause; Jennifer Dy. Vol. 3 International Machine Learning Society (IMLS), 2018. p. 1726-1734.

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

Cortes, C, Desalvo, G, Gentile, C, Mohri, M & Yang, S 2018, Online Learning with Abstention. in A Krause & J Dy (eds), 35th International Conference on Machine Learning, ICML 2018. vol. 3, International Machine Learning Society (IMLS), pp. 1726-1734, 35th International Conference on Machine Learning, ICML 2018, Stockholm, Sweden, 7/10/18.
Cortes C, Desalvo G, Gentile C, Mohri M, Yang S. Online Learning with Abstention. In Krause A, Dy J, editors, 35th International Conference on Machine Learning, ICML 2018. Vol. 3. International Machine Learning Society (IMLS). 2018. p. 1726-1734
Cortes, Corinna ; Desalvo, Giulia ; Gentile, Claudio ; Mohri, Mehryar ; Yang, Scott. / Online Learning with Abstention. 35th International Conference on Machine Learning, ICML 2018. editor / Andreas Krause ; Jennifer Dy. Vol. 3 International Machine Learning Society (IMLS), 2018. pp. 1726-1734
@inproceedings{5ab6f2d15bd1485186975f3b34293370,
title = "Online Learning with Abstention",
abstract = "We present an extensive study of a key problem in online learning where the learner can opt to abstain from making a prediction, at a ccrtain cost. In the adversarial setting, we show how existing online algorithms and guarantees can be adapted to this problem. In the stochastic setting, we first point out a bias problem that limits the straightforward extension of algorithms such as ucb-n to this context. Next, we give a new algorithm, ucb- gt, that exploits historical data and time-varying feedback graphs. We show that this algorithm ben: cfits from more favorable regret guarantees than a natural extension of ucb-n. We further report the results of a series of experiments demonstrating that ucb-gt largely outperforms that extension of ucb-n, as well as other standard baselines.",
author = "Corinna Cortes and Giulia Desalvo and Claudio Gentile and Mehryar Mohri and Scott Yang",
year = "2018",
month = "1",
day = "1",
language = "English (US)",
volume = "3",
pages = "1726--1734",
editor = "Andreas Krause and Jennifer Dy",
booktitle = "35th International Conference on Machine Learning, ICML 2018",
publisher = "International Machine Learning Society (IMLS)",

}

TY - GEN

T1 - Online Learning with Abstention

AU - Cortes, Corinna

AU - Desalvo, Giulia

AU - Gentile, Claudio

AU - Mohri, Mehryar

AU - Yang, Scott

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We present an extensive study of a key problem in online learning where the learner can opt to abstain from making a prediction, at a ccrtain cost. In the adversarial setting, we show how existing online algorithms and guarantees can be adapted to this problem. In the stochastic setting, we first point out a bias problem that limits the straightforward extension of algorithms such as ucb-n to this context. Next, we give a new algorithm, ucb- gt, that exploits historical data and time-varying feedback graphs. We show that this algorithm ben: cfits from more favorable regret guarantees than a natural extension of ucb-n. We further report the results of a series of experiments demonstrating that ucb-gt largely outperforms that extension of ucb-n, as well as other standard baselines.

AB - We present an extensive study of a key problem in online learning where the learner can opt to abstain from making a prediction, at a ccrtain cost. In the adversarial setting, we show how existing online algorithms and guarantees can be adapted to this problem. In the stochastic setting, we first point out a bias problem that limits the straightforward extension of algorithms such as ucb-n to this context. Next, we give a new algorithm, ucb- gt, that exploits historical data and time-varying feedback graphs. We show that this algorithm ben: cfits from more favorable regret guarantees than a natural extension of ucb-n. We further report the results of a series of experiments demonstrating that ucb-gt largely outperforms that extension of ucb-n, as well as other standard baselines.

UR - http://www.scopus.com/inward/record.url?scp=85057257985&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057257985&partnerID=8YFLogxK

M3 - Conference contribution

VL - 3

SP - 1726

EP - 1734

BT - 35th International Conference on Machine Learning, ICML 2018

A2 - Krause, Andreas

A2 - Dy, Jennifer

PB - International Machine Learning Society (IMLS)

ER -