Advertisement

Transfer Learning through Domain Adaptation

  • Huaxiang Zhang
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6677)

Abstract

It is interesting and helpful to use the labeled data of some tasks to improve the classification performance of another task. This paper focuses on this issue and proposes an algorithm named SSDT (Synthetic Source Data Transfer). As the number of the training data influences the classification performance greatly, we create some synthetic training data using the source data and combine them with the target data to train a classifier. The classifier is applied to the target data, and experimental results show that SSDT improves the performance obviously.

Keywords

KNN data distribution transfer learning classification 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Pan, S.J., Yang, Q.: A Survey on Transfer Learning (2008), http://www.cs.ust.hk/~qyang/publications.htm
  2. 2.
    Bickel, S., Bruckner, M., Scheffer, T.: Discriminative Learning for Differing Training and Test Distributions. In: 24th International Conference on Machine Learning, New York, pp. 81–88 (2007)Google Scholar
  3. 3.
    Dai, W.Y., Yang, Q., Xue, G.R., Yu, Y.: Boosting for transfer learning. In: 24th International Conference on Machine Learning, Corvallis, pp. 193–200 (2007)Google Scholar
  4. 4.
    Jiang, J., Zhai, C.X.: Instance Weighting for Domain Adaptation in NLP. In: 45th Annual Meeting of the Association of Computational Linguistics, Prague, pp. 264–271 (2007)Google Scholar
  5. 5.
    Raina, R., Battle, A., Lee, H., Packer, B., Andrew, Y.N.: Self-Taught Learning: Transfer Learning from Unlabeled Data. In: 24th International Conferenceon Machine Learning, Corvallis, pp. 759–766 (2007)Google Scholar
  6. 6.
    Lee, S.I., Chatalbashev, V., Vickrey, D., Koller, D.: Learning a Meta-Level Prior for Feature Relevance from Multiple Related Tasks. In: 24th International Conference on Machine Learning, Corvalis, pp. 489–496 (2007)Google Scholar
  7. 7.
    Daume III, H.: Frustratingly Easy Domain Adaptation. In: 45th Annual Meeting of the Association of Computational Linguistics, Prague, pp. 256–263 (2007)Google Scholar
  8. 8.
    Blitzer, J., McDonald, R., Pereira, F.: Domain Adaptation with Structural Correspondence Learning. In: Conference on Empirical Methods in Natural Language, Sydney, pp. 120–128 (2006)Google Scholar
  9. 9.
    Argyriou, A., Evgeniou, T., Pontil, M.: Multi-Task Feature Learning. In: 19th Annual Conference on Neural Information Processing Systems, Vancouver, pp. 41–48 (2007)Google Scholar
  10. 10.
    Schwaighofer, A., Tresp, V., Yu, K.: Learning Gaussian process kernels via hierarchical bayes. In: 17th Annual Conference on Neural Information Processing Systems, pp. 1209–1216. MIT Press, Cambridge (2005)Google Scholar
  11. 11.
    Bonilla, E., Chai, K.M., Williams, C.: Multi-task Gaussian Process Prediction. In: 20th Annual Conference on Neural Information Processing Systems, pp. 153–160. MIT Press, Cambridge (2008)Google Scholar
  12. 12.
    Daume III, H., Marcu, D.: for Statistical Classifiers. Journal of Artificial Intelligence Research 26, 101–126 (2006)Google Scholar
  13. 13.
    Chuong, B.D., Andrew, Y.N.: Transfer Learning for Text Classification. Journal of Advances in Neural Information Processing Systems 18, 299–306 (2006)Google Scholar
  14. 14.
    Dai, W.Y., Xue, G.R., Yang, Q., Yu, Y.: Co-clustering Based Classification for out-of-domain Documents. In: KDD, California, pp. 201–219 (2007)Google Scholar
  15. 15.
    Wu, P.C., Dietterich, T.G.: Improving SVM Accuracy by Training on Auxiliary Data Sources. In: 21st International Conference on Machine Learning, Banff (2004)Google Scholar
  16. 16.
    Chawla, N.V., Hall, L.O., Bowyer, K.W., Kegelmeyer, W.P.: SMOTE: Synthetic Minority Oversampling Technique. J. Artificial Intelligence Research 16, 321–357 (2002)zbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Huaxiang Zhang
    • 1
  1. 1.Department of Computer ScienceShandong Normal UniversityJinanChina

Personalised recommendations