Does unlabeled data provably help? Worst-case analysis of the sample complexity of semi-supervised learning

21st Annual Conference on Learning Theory, COLT 2008

Volumn , Issue , 2008, Pages 33-44

  Ben David, Shai ^a   Lu, Tyler ^a   Pál, Dávid ^a  

^a UNIVERSITY OF WATERLOO   (Canada)

Author keywords

[No Author keywords available]

Indexed keywords

CLASSIFICATION PREDICTION; CLUSTER ASSUMPTIONS; DISTRIBUTION-FREE; POTENTIAL BENEFITS; PREDICTION PERFORMANCE; SAMPLE COMPLEXITY; SEMI-SUPERVISED LEARNING; WORST-CASE ANALYSIS;

SUPERVISED LEARNING;

EID: 77956501439     PISSN: None     EISSN: None     Source Type: Conference Proceeding    
DOI: None     Document Type: Conference Paper

References (15)

1. Dana Angluin and Philip D. Laird. Learning from noisy examples. Machine Learning, 2(4): 343-370, 1987.
2. Martin Anthony and Peter L. Bartlett. Neural Network Learning: Theoretical Foundations. Cambridge University Press, January 1999.
3. Maria-Florina Balcan and Avrim Blum. A PAC-style model for learning from labeled and unlabeled data. In Proceedings of 18th Annual Conference on Learning Theory 2005, pages 111-126. Springer, 2005.
4. Maria-Florina Balcan and Avrim Blum. An augmented PAC model for semi-supervised learning. In Olivier Chapelle, Bernhard Schölkopf, and Alexander Zien, editors, Semi-Supervised Learning, chapter 21, pages 61-89. MIT Press, September 2006.
5. Gyora M. Benedek and Alon Itai. Learnability with respect to fixed distributions. Theoretical Computer Science, 86(2): 377-389, 1991.
6. Avrim Blum and Tom M. Mitchell. Combining labeled and unlabeled sata with co-training. In COLT, pages 92-100, 1998.
7. Olivier Chapelle, Bernhard Schölkopf, and Alexander Zien, editors. Semi-Supervised Learning. MIT Press, September 2006.
8. Fabio Cozman and Ira Cohen. Risks of semi-supervised learning: How unlabeled data can degrade performance of generative classifiers. In Olivier Chapelle, Bernhard Schölkopf, and Alexander Zien, editors, Semi-Supervised Learning, chapter 4, pages 57-72. MIT Press, September 2006.
9. Ran El-Yaniv and Dmitry Pechyony. Stable transductive learning. In COLT, pages 35-49, 2006.
10. Ran El-Yaniv and Dmitry Pechyony. Transductive rademacher complexity and its applications. In COLT, pages 157-171, 2007.
11. Claudio Gentile and David P. Helmbold. Improved lower bounds for learning from noisy examples: and information-theoretic approach. In Proceedings of COLT 1998, pages 104-115. ACM, 1998.
12. Matti Kääriäinen. Generalization error bounds using unlabeled data. In Proceedings of COLT 2005, pages 127-142. Springer, 2005.
13. P. Rigollet. Generalization error bounds in semi-supervised classification under the cluster assumption. The Journal of Machine Learning Research, 8: 1369-1392, 2007.
14. Vladimir N. Vapnik. Statistical Learning Theory. Wiley-Interscience, September 1998.
15. Vladimir N. Vapnik. Transductive inference and semi-supervised learning. In Olivier Chapelle, Bernhard Schölkopf, and Alexander Zien, editors, Semi-Supervised Learning, chapter 24, pages 453-472. MIT Press, September 2006.