Abstract
Recent years have witnessed an increasing interest in transfer learning. This paper deals with the classification problem that the target-domain with a different distribution from the source-domain is totally unlabeled, and aims to build an inductive model for unseen data. Firstly, we analyze the problem of class ratio drift in the previous work of transductive transfer learning, and propose to use a normalization method to move towards the desired class ratio. Furthermore, we develop a hybrid regularization framework for inductive transfer learning. It considers three factors, including the distribution geometry of the target-domain by manifold regularization, the entropy value of prediction probability by entropy regularization, and the class prior by expectation regularization. This framework is used to adapt the inductive model learnt from the source-domain to the target-domain. Finally, the experiments on the real-world text data show the effectiveness of our inductive method of transfer learning. Meanwhile, it can handle unseen test points.
Similar content being viewed by others
References
Dai W Y, Yang Q, Xue G R, et al. Boosting for Transfer Learning. In: Ghahramani Z B, eds. Proceeding of 24th International Conference on Machine Learning, 2007 Jun 20–24, Corvalis, Oregon. ACM, 2007. 193–200
Raina R, Battle A, Lee H, et al. Self-taught Learning: Transfer Learning from Unlabeled Data. In: Ghahramani Z B, eds. Proceeding of 24th International Conference on Machine Learning, 2007 Jun 20–24, Corvalis, Oregon. ACM, 2007. 759–766
Dai W Y, Xue G R, Yang Q, et al. Co-clustering based Classification for Out-of-domain Documents. In: Berkhin P, Caruana R, Wu X D, et al, eds. Proceeding of 13th ACM International Conference on Knowledge Discovery and Data Mining, 2007, Aug 12–15, San Jose, California. ACM, 2007. 210–219
Xing D K, Dai W Y, Xue G R, et al. Bridged Refinement for Transfer Learning. In: Joost N K, Jacek K, Ramon L, et al, eds. Proceeding of 11th European Conference on Practice of Knowledge Discovery in Databases, 2007 Sep 17–21, Warsaw Poland. Springer, 2007. 324–335
Liao X J, Xue Y, Carin L, et al. Logistic Regression with an Auxiliary Data Source. In: Raedt L D, Wrobel S, eds. Proceeding of 22th International Conference on Machine Learning, 2007 Aug 7–11, Bonn, Germany. ACM, 2005. 505–512
Yang J, Yan R, Hauptmann A G, et al. Cross-domain Video Concept Detection Using Adaptive SVMs. In Rainer L, Anand R P, Alan H, et al, eds. Proceeding of 15th International Conference on Multimedia, 2007 Sep 24–29, Augsburg Germany. ACM, 2007. 188–197
Wu P C, Dietterich T G, et al. Improving SVM Accuracy by Training on Auxiliary Data Sources. In: Brodley C E, eds. Proceeding of 21th International Conference on Machine Learning, 2004 Jul 4–8, Banff, Alberta, Canada. ACM, 2004. 871–878
Mahmud M M H, Ray S, et al. Transfer Learning Using Kolmogorov Complexity: Basic Theory and Empirical Evaluations. Technical Report, UIUC-DCS-R-2007-2875, Department of Computer Science, University of Illinois at Urbana-Champaign. 2007
Ben-David S, Blitzer J, Crammer K, Pereira F, et al. Analysis of Representations for Domain Adaptation. In: Koller D, Singer Y, Platt J, et al, eds. Proceeding of Advances in Neural Information Processing Systems 20, MIT Press, Cambridge, MA, 2007(20): 137–144
Dai W Y, Xue G R, Yang Q, et al. Transferring Naive Bayes Classifiers for Text Classification. In: John C, Peyman F, Simon P, et al, eds. Proceeding of 22nd Conference on Artificial Intelligence, 2007 Jul 22–26, Vancouver, British Columbia. AAAI Press, 2007. 540–545
Samarth S, Sylvian R, et al. Cross Domain Knowledge Transfer Using Structured Representations. In: Proceeding of 21nd Conference on Artificial Intelligence, 2006 Jul 16–22, Boston, Massachusetts. AAAI Press, 2006
Ling X, Dai W Y, Xue G R, et al. Spectral Domain-Transfer Learning. In: Li Y, Liu B, Sunita S, et al, eds. Proceeding of 14th ACM International Conference on Knowledge Discovery and Data Mining, 2008, Aug 24–27, Las Vegas, Nevada. ACM, 2008. 488–496
Mahmud, M M H. On Universal Transfer Learning. In: Rocco M H, Servedio R A, Takimoto E, et al, eds. Proceeding of 18th International Conference on Algorithmic Learning Theory, 2007, Oct 1–4, Sendai, Japan. LNCS, 2007. 135–149
Belkin M, Niyogi P, Sindhwani V, et al. Manifold Regularization: A Geometric Framework for Learning from Labeled and Unlabeled Examples. Journal of Machine Learning Research, 2006: 2399–2434
Grandvalet Y, Bengio Y, et al. Semi-supervised Learning by Entropy Minimization. Proceeding of 19th Conference on Neural Information Processing Systems, 2005 Dec 5–8, Vancouver, British Columbia. MIT Press, 2005. 529–536
Mann G S, McCallum A, et al. Simple, Robust, Scalable Semi-supervised Learning via Expectation Regularization. In: Ghahramani Z B, eds. Proceeding of 24th International Conference on Machine Learning, 2007 Jun 20–24, Corvalis, Oregon. ACM, 2007. 593–600
Wang F, Zhang C S, et al. Label Propagation through Linear Neighborhoods. IEEE Transations on Knowledge and Data Engineering, 2008: 55–67
Joachims T. Making Large-scale SVM Learning Practical. In: Scholkopf B, Burges C, Smola A, et al, eds. Proceeding of Advances in Kernel Methods, 1999. MIT Press, Cambridge, 1999. 169–184
Davie H, Stanley L, et al. Applied Logistic Regression. Wiley, New York, 2000
Joachims T. Transductive Inference for Text Classification Using Support Vector Machines. In: Bratko I, Dzeroski S, eds. Proceeding of 16th International Conference on Machine Learning, 1999 Jun 27–30, Bled, Slovenia. ACM, 1999. 200–209
Joachims T. Transductive Learning via Spectral Graph Partitioning. In: Tom F, Nina M, eds. Proceeding of 20th International Conference on Machine Learning, 2003 Aug 21–24, Washington DC. ACM, 2003. 290–297
Author information
Authors and Affiliations
Corresponding author
Additional information
Supported by the National Science Foundation of China (Grant Nos. 60435010, 60675010), National High Technology Research and Development of China (Grant Nos. 2006AA01Z128, 2007AA01Z132), National Basic Research Priorities Programme (Grant No. 2007CB311004) and National Science and Technology Support Plan (Grant No. 2006BAC08B06)
Electronic supplementary material
About this article
Cite this article
Zhuang, F., Luo, P., He, Q. et al. Inductive transfer learning for unlabeled target-domain via hybrid regularization. Chin. Sci. Bull. 54, 2470–2478 (2009). https://doi.org/10.1007/s11434-009-0171-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11434-009-0171-x