Dictionary based sparse representation for domain adaptation

Rishabh Mehrotra, Rushabh Agrawal, Syed Aqueel Haider

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

9 Citations (Scopus)

Abstract

Machine Learning algorithms are often as good as the data they can learn from. Enormous amount of unlabeled data is readily available and the ability to efficiently use such amount of unlabeled data holds a significant promise in terms of increasing the performance of various learning tasks. We consider the task of supervised Domain Adaptation and present a Self-Taught learning based framework which makes use of the K-SVD algorithm for learning sparse representation of data in an unsupervised manner. To the best of our knowledge this is the first work that integrates K-SVD algorithm into the self-taught learning framework. The K-SVD algorithm iteratively alternates between sparse coding of the instances based on the current dictionary and a process of updating/adapting the dictionary to better fit the data so as to achieve a sparse representation under strict sparsity constraints. Using the learnt dictionary, a rich feature representation of the few labeled instances is obtained which is fed to a classifier along with class labels to build the model. We evaluate our framework on the task of domain adaptation for sentiment classification. Both self-domain (requiring very few domain-specific training instances) and cross-domain classification (requiring 0 labeled instances of target domain and very few labeled instances of source domain) are performed. Empirical comparisons of self-domain and cross-domain results establish the efficacy of the proposed framework.

Original languageEnglish
Title of host publicationCIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management
Pages2395-2398
Number of pages4
DOIs
Publication statusPublished - 19-12-2012
Externally publishedYes
Event21st ACM International Conference on Information and Knowledge Management, CIKM 2012 - Maui, HI, United States
Duration: 29-10-201202-11-2012

Conference

Conference21st ACM International Conference on Information and Knowledge Management, CIKM 2012
CountryUnited States
CityMaui, HI
Period29-10-1202-11-12

Fingerprint

Singular value decomposition
Glossaries
Learning algorithms
Learning systems
Labels
Classifiers

All Science Journal Classification (ASJC) codes

  • Human-Computer Interaction
  • Computer Networks and Communications
  • Computer Vision and Pattern Recognition
  • Software

Cite this

Mehrotra, R., Agrawal, R., & Haider, S. A. (2012). Dictionary based sparse representation for domain adaptation. In CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management (pp. 2395-2398) https://doi.org/10.1145/2396761.2398649
Mehrotra, Rishabh ; Agrawal, Rushabh ; Haider, Syed Aqueel. / Dictionary based sparse representation for domain adaptation. CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management. 2012. pp. 2395-2398
@inproceedings{98bf9d0d478a480190f72cc5b40540f9,
title = "Dictionary based sparse representation for domain adaptation",
abstract = "Machine Learning algorithms are often as good as the data they can learn from. Enormous amount of unlabeled data is readily available and the ability to efficiently use such amount of unlabeled data holds a significant promise in terms of increasing the performance of various learning tasks. We consider the task of supervised Domain Adaptation and present a Self-Taught learning based framework which makes use of the K-SVD algorithm for learning sparse representation of data in an unsupervised manner. To the best of our knowledge this is the first work that integrates K-SVD algorithm into the self-taught learning framework. The K-SVD algorithm iteratively alternates between sparse coding of the instances based on the current dictionary and a process of updating/adapting the dictionary to better fit the data so as to achieve a sparse representation under strict sparsity constraints. Using the learnt dictionary, a rich feature representation of the few labeled instances is obtained which is fed to a classifier along with class labels to build the model. We evaluate our framework on the task of domain adaptation for sentiment classification. Both self-domain (requiring very few domain-specific training instances) and cross-domain classification (requiring 0 labeled instances of target domain and very few labeled instances of source domain) are performed. Empirical comparisons of self-domain and cross-domain results establish the efficacy of the proposed framework.",
author = "Rishabh Mehrotra and Rushabh Agrawal and Haider, {Syed Aqueel}",
year = "2012",
month = "12",
day = "19",
doi = "10.1145/2396761.2398649",
language = "English",
isbn = "9781450311564",
pages = "2395--2398",
booktitle = "CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management",

}

Mehrotra, R, Agrawal, R & Haider, SA 2012, Dictionary based sparse representation for domain adaptation. in CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management. pp. 2395-2398, 21st ACM International Conference on Information and Knowledge Management, CIKM 2012, Maui, HI, United States, 29-10-12. https://doi.org/10.1145/2396761.2398649

Dictionary based sparse representation for domain adaptation. / Mehrotra, Rishabh; Agrawal, Rushabh; Haider, Syed Aqueel.

CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management. 2012. p. 2395-2398.

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

TY - GEN

T1 - Dictionary based sparse representation for domain adaptation

AU - Mehrotra, Rishabh

AU - Agrawal, Rushabh

AU - Haider, Syed Aqueel

PY - 2012/12/19

Y1 - 2012/12/19

N2 - Machine Learning algorithms are often as good as the data they can learn from. Enormous amount of unlabeled data is readily available and the ability to efficiently use such amount of unlabeled data holds a significant promise in terms of increasing the performance of various learning tasks. We consider the task of supervised Domain Adaptation and present a Self-Taught learning based framework which makes use of the K-SVD algorithm for learning sparse representation of data in an unsupervised manner. To the best of our knowledge this is the first work that integrates K-SVD algorithm into the self-taught learning framework. The K-SVD algorithm iteratively alternates between sparse coding of the instances based on the current dictionary and a process of updating/adapting the dictionary to better fit the data so as to achieve a sparse representation under strict sparsity constraints. Using the learnt dictionary, a rich feature representation of the few labeled instances is obtained which is fed to a classifier along with class labels to build the model. We evaluate our framework on the task of domain adaptation for sentiment classification. Both self-domain (requiring very few domain-specific training instances) and cross-domain classification (requiring 0 labeled instances of target domain and very few labeled instances of source domain) are performed. Empirical comparisons of self-domain and cross-domain results establish the efficacy of the proposed framework.

AB - Machine Learning algorithms are often as good as the data they can learn from. Enormous amount of unlabeled data is readily available and the ability to efficiently use such amount of unlabeled data holds a significant promise in terms of increasing the performance of various learning tasks. We consider the task of supervised Domain Adaptation and present a Self-Taught learning based framework which makes use of the K-SVD algorithm for learning sparse representation of data in an unsupervised manner. To the best of our knowledge this is the first work that integrates K-SVD algorithm into the self-taught learning framework. The K-SVD algorithm iteratively alternates between sparse coding of the instances based on the current dictionary and a process of updating/adapting the dictionary to better fit the data so as to achieve a sparse representation under strict sparsity constraints. Using the learnt dictionary, a rich feature representation of the few labeled instances is obtained which is fed to a classifier along with class labels to build the model. We evaluate our framework on the task of domain adaptation for sentiment classification. Both self-domain (requiring very few domain-specific training instances) and cross-domain classification (requiring 0 labeled instances of target domain and very few labeled instances of source domain) are performed. Empirical comparisons of self-domain and cross-domain results establish the efficacy of the proposed framework.

UR - http://www.scopus.com/inward/record.url?scp=84871059356&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871059356&partnerID=8YFLogxK

U2 - 10.1145/2396761.2398649

DO - 10.1145/2396761.2398649

M3 - Conference contribution

AN - SCOPUS:84871059356

SN - 9781450311564

SP - 2395

EP - 2398

BT - CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management

ER -

Mehrotra R, Agrawal R, Haider SA. Dictionary based sparse representation for domain adaptation. In CIKM 2012 - Proceedings of the 21st ACM International Conference on Information and Knowledge Management. 2012. p. 2395-2398 https://doi.org/10.1145/2396761.2398649