Hierarchical active transfer learning

David Kale, Marjan Ghazvininejad, Anil Ramakrishna, Jingrui He, Yan Liu

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

12 Citations (Scopus)

Abstract

We describe a unified active transfer learning framework called Hierarchical Active Transfer Learning (HATL). HATL exploits cluster structure shared between different data domains to perform transfer learning by imputing labels for unlabeled target data and to generate effective label queries during active learning. The resulting framework is flexible enough to perform not only adaptive transfer learning and accelerated active learning but also unsupervised and semi-supervised transfer learning. We derive an intuitive and useful upper bound on HATL's error when used to infer labels for unlabeled target points. We also present results on synthetic data that confirm both intuition and our analysis. Finally, we demonstrate HATL's empirical effectiveness on a benchmark data set for sentiment classification.

Original languageEnglish (US)
Title of host publicationSIAM International Conference on Data Mining 2015, SDM 2015
PublisherSociety for Industrial and Applied Mathematics Publications
Pages514-522
Number of pages9
ISBN (Print)9781510811522
StatePublished - 2015
EventSIAM International Conference on Data Mining 2015, SDM 2015 - Vancouver, Canada
Duration: Apr 30 2015May 2 2015

Other

OtherSIAM International Conference on Data Mining 2015, SDM 2015
CountryCanada
CityVancouver
Period4/30/155/2/15

Fingerprint

Labels
Problem-Based Learning

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Computer Vision and Pattern Recognition
  • Software

Cite this

Kale, D., Ghazvininejad, M., Ramakrishna, A., He, J., & Liu, Y. (2015). Hierarchical active transfer learning. In SIAM International Conference on Data Mining 2015, SDM 2015 (pp. 514-522). Society for Industrial and Applied Mathematics Publications.

Hierarchical active transfer learning. / Kale, David; Ghazvininejad, Marjan; Ramakrishna, Anil; He, Jingrui; Liu, Yan.

SIAM International Conference on Data Mining 2015, SDM 2015. Society for Industrial and Applied Mathematics Publications, 2015. p. 514-522.

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

Kale, D, Ghazvininejad, M, Ramakrishna, A, He, J & Liu, Y 2015, Hierarchical active transfer learning. in SIAM International Conference on Data Mining 2015, SDM 2015. Society for Industrial and Applied Mathematics Publications, pp. 514-522, SIAM International Conference on Data Mining 2015, SDM 2015, Vancouver, Canada, 4/30/15.
Kale D, Ghazvininejad M, Ramakrishna A, He J, Liu Y. Hierarchical active transfer learning. In SIAM International Conference on Data Mining 2015, SDM 2015. Society for Industrial and Applied Mathematics Publications. 2015. p. 514-522
Kale, David ; Ghazvininejad, Marjan ; Ramakrishna, Anil ; He, Jingrui ; Liu, Yan. / Hierarchical active transfer learning. SIAM International Conference on Data Mining 2015, SDM 2015. Society for Industrial and Applied Mathematics Publications, 2015. pp. 514-522
@inproceedings{e504e8d907e4410485a70b20762851d7,
title = "Hierarchical active transfer learning",
abstract = "We describe a unified active transfer learning framework called Hierarchical Active Transfer Learning (HATL). HATL exploits cluster structure shared between different data domains to perform transfer learning by imputing labels for unlabeled target data and to generate effective label queries during active learning. The resulting framework is flexible enough to perform not only adaptive transfer learning and accelerated active learning but also unsupervised and semi-supervised transfer learning. We derive an intuitive and useful upper bound on HATL's error when used to infer labels for unlabeled target points. We also present results on synthetic data that confirm both intuition and our analysis. Finally, we demonstrate HATL's empirical effectiveness on a benchmark data set for sentiment classification.",
author = "David Kale and Marjan Ghazvininejad and Anil Ramakrishna and Jingrui He and Yan Liu",
year = "2015",
language = "English (US)",
isbn = "9781510811522",
pages = "514--522",
booktitle = "SIAM International Conference on Data Mining 2015, SDM 2015",
publisher = "Society for Industrial and Applied Mathematics Publications",

}

TY - GEN

T1 - Hierarchical active transfer learning

AU - Kale, David

AU - Ghazvininejad, Marjan

AU - Ramakrishna, Anil

AU - He, Jingrui

AU - Liu, Yan

PY - 2015

Y1 - 2015

N2 - We describe a unified active transfer learning framework called Hierarchical Active Transfer Learning (HATL). HATL exploits cluster structure shared between different data domains to perform transfer learning by imputing labels for unlabeled target data and to generate effective label queries during active learning. The resulting framework is flexible enough to perform not only adaptive transfer learning and accelerated active learning but also unsupervised and semi-supervised transfer learning. We derive an intuitive and useful upper bound on HATL's error when used to infer labels for unlabeled target points. We also present results on synthetic data that confirm both intuition and our analysis. Finally, we demonstrate HATL's empirical effectiveness on a benchmark data set for sentiment classification.

AB - We describe a unified active transfer learning framework called Hierarchical Active Transfer Learning (HATL). HATL exploits cluster structure shared between different data domains to perform transfer learning by imputing labels for unlabeled target data and to generate effective label queries during active learning. The resulting framework is flexible enough to perform not only adaptive transfer learning and accelerated active learning but also unsupervised and semi-supervised transfer learning. We derive an intuitive and useful upper bound on HATL's error when used to infer labels for unlabeled target points. We also present results on synthetic data that confirm both intuition and our analysis. Finally, we demonstrate HATL's empirical effectiveness on a benchmark data set for sentiment classification.

UR - http://www.scopus.com/inward/record.url?scp=84961943273&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961943273&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781510811522

SP - 514

EP - 522

BT - SIAM International Conference on Data Mining 2015, SDM 2015

PB - Society for Industrial and Applied Mathematics Publications

ER -