Encyclopedia of Algorithms

2016 Edition
| Editors: Ming-Yang Kao

Semi-supervised Learning

  • Avrim Blum
Reference work entry
DOI: https://doi.org/10.1007/978-1-4939-2864-4_766

Years and Authors of Summarized Original Work

  • 1998; Blum, Mitchell

  • 1999; Joachims

  • 2010; Balcan, Blum

Problem Definition

Semi-supervised learning [1, 4, 5, 8, 12] refers to the problem of using a large unlabeled data set U together with a given labeled data set L in order to generate prediction rules that are more accurate on new data than would have been achieved using just L alone. Semi-supervised learning is motivated by the fact that in many settings (e.g., document classification, image classification, speech recognition), unlabeled data is plentiful but labeled data is more limited or expensive, e.g., due to the need for human labelers. Therefore, one would like to make use of the unlabeled data if possible.

The general idea behind semi-supervised learning is that unlabeled data, while missing the labels, nonetheless often contains useful information. As an example, suppose one believes the correct decision boundary for some classification problem should be a linear separator that...

Keywords

Co-training Learning from labeled and unlabeled data Semi-supervised SVM 
This is a preview of subscription content, log in to check access.

Recommended Reading

  1. 1.
    Balcan MF, Blum A (2010) A discriminative model for semi-supervised learning. J ACM 57(3):19:1–19:46. doi:10.1145/1706591.1706599. http://doi.acm.org/10.1145/1706591.1706599
  2. 2.
    Balcan MF, Blum A, Yang K (2004) Co-training and expansion: towards bridging theory and practice. In: Proceedings of 18th conference on neural information processing systems, VancouverGoogle Scholar
  3. 3.
    Blum A, Chawla S (2001) Learning from labeled and unlabeled data using graph mincuts. In: Proceedings of 18th international conference on machine learning, Williams CollegeGoogle Scholar
  4. 4.
    Blum A, Mitchell TM (1998) Combining labeled and unlabeled data with co-training. In: Proceedings of the 11th annual conference on computational learning theory, Madison, pp 92– 100Google Scholar
  5. 5.
    Chapelle O, Schölkopf B, Zien A (eds) (2006) Semi-supervised learning. MIT, Cambridge. http://www.kyb.tuebingen.mpg.de/ssl-book
  6. 6.
    Collins M, Singer Y (1999) Unsupervised models for named entity classification. In: Proceedings of the joint SIGDAT conference on empirical methods in natural language processing and very large corpora, College Park, pp 189–196Google Scholar
  7. 7.
    Gupta S, Kim J, Grauman K, Mooney R (2008) Watch, listen & learn: co-training on captioned images and videos. In: Machine learning and knowledge discovery in databases (ECML PKDD). Lecture notes in computer science, vol 5211. Springer, Berlin/Heidelberg, pp 457–472. 10.1007/978-3-540-87479-9_48. http://dx.doi.org/10.1007/978-3-540-87479-9_48
  8. 8.
    Joachims T (1999) Transductive inference for text classification using support vector machines. In: Proceedings of 16th international conference on machine learning, Bled, pp 200–209Google Scholar
  9. 9.
    Levin A, Viola P, Freund Y (2003) Unsupervised improvement of visual detectors using co-training. In: Proceedings of the ninth IEEE international conference on computer vision, ICCV ’03, vol 2, Nice. IEEE Computer Society, Washington, DC, pp 626–633. http://dl.acm.org/citation.cfm?id=946247.946615
  10. 10.
    Nigam K, Ghani R (2000) Analyzing the effectiveness and applicability of co-training. In: Proceedings of ACM CIKM international conference on information and knowledge management, McLean, pp 86–93Google Scholar
  11. 11.
    Vapnik V (1998) Statistical learning theory, vol 2. Wiley, New YorkzbMATHGoogle Scholar
  12. 12.
    Zhu X (2006) Semi-supervised learning literature survey Computer sciences TR 1530 University of Wisconsin, MadisonGoogle Scholar
  13. 13.
    Zhu X, Ghahramani Z, Lafferty J (2003) Semi-supervised learning using Gaussian fields and harmonic functions. In: Proceedings of 20th international conference on machine learning, Washington, DC, pp 912–919Google Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Avrim Blum
    • 1
  1. 1.School of Computer ScienceCarnegie Mellon UniversityPittsburghUSA