Scalable Heterogeneous Transfer Ranking

Mohammad Taha Bahadori, Yi Chang, Bo Long, Yan Liu
Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications, PMLR 36:214-228, 2014.

Abstract

In this paper, we propose to study the problem of heterogeneous transfer ranking, a transfer learning problem with heterogeneous features in order to utilize the rich large-scale labeled data in popular languages to help the ranking task in less popular languages. We develop a large-margin algorithm, namely LM-HTR, to solve the problem by mapping the input features in both the source domain and target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We analyze the theoretical bound of the prediction loss and develop fast algorithms via stochastic gradient descent so that our model can be scalable to large-scale applications. Experiment results on two application datasets demonstrate the advantages of our algorithms over other state-of-the-art methods.

Cite this Paper


BibTeX
@InProceedings{pmlr-v36-bahadori14, title = {Scalable Heterogeneous Transfer Ranking}, author = {Bahadori, Mohammad Taha and Chang, Yi and Long, Bo and Liu, Yan}, booktitle = {Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications}, pages = {214--228}, year = {2014}, editor = {Fan, Wei and Bifet, Albert and Yang, Qiang and Yu, Philip S.}, volume = {36}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {24 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v36/bahadori14.pdf}, url = {https://proceedings.mlr.press/v36/bahadori14.html}, abstract = {In this paper, we propose to study the problem of heterogeneous transfer ranking, a transfer learning problem with heterogeneous features in order to utilize the rich large-scale labeled data in popular languages to help the ranking task in less popular languages. We develop a large-margin algorithm, namely LM-HTR, to solve the problem by mapping the input features in both the source domain and target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We analyze the theoretical bound of the prediction loss and develop fast algorithms via stochastic gradient descent so that our model can be scalable to large-scale applications. Experiment results on two application datasets demonstrate the advantages of our algorithms over other state-of-the-art methods.} }
Endnote
%0 Conference Paper %T Scalable Heterogeneous Transfer Ranking %A Mohammad Taha Bahadori %A Yi Chang %A Bo Long %A Yan Liu %B Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications %C Proceedings of Machine Learning Research %D 2014 %E Wei Fan %E Albert Bifet %E Qiang Yang %E Philip S. Yu %F pmlr-v36-bahadori14 %I PMLR %P 214--228 %U https://proceedings.mlr.press/v36/bahadori14.html %V 36 %X In this paper, we propose to study the problem of heterogeneous transfer ranking, a transfer learning problem with heterogeneous features in order to utilize the rich large-scale labeled data in popular languages to help the ranking task in less popular languages. We develop a large-margin algorithm, namely LM-HTR, to solve the problem by mapping the input features in both the source domain and target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We analyze the theoretical bound of the prediction loss and develop fast algorithms via stochastic gradient descent so that our model can be scalable to large-scale applications. Experiment results on two application datasets demonstrate the advantages of our algorithms over other state-of-the-art methods.
RIS
TY - CPAPER TI - Scalable Heterogeneous Transfer Ranking AU - Mohammad Taha Bahadori AU - Yi Chang AU - Bo Long AU - Yan Liu BT - Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications DA - 2014/08/13 ED - Wei Fan ED - Albert Bifet ED - Qiang Yang ED - Philip S. Yu ID - pmlr-v36-bahadori14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 36 SP - 214 EP - 228 L1 - http://proceedings.mlr.press/v36/bahadori14.pdf UR - https://proceedings.mlr.press/v36/bahadori14.html AB - In this paper, we propose to study the problem of heterogeneous transfer ranking, a transfer learning problem with heterogeneous features in order to utilize the rich large-scale labeled data in popular languages to help the ranking task in less popular languages. We develop a large-margin algorithm, namely LM-HTR, to solve the problem by mapping the input features in both the source domain and target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We analyze the theoretical bound of the prediction loss and develop fast algorithms via stochastic gradient descent so that our model can be scalable to large-scale applications. Experiment results on two application datasets demonstrate the advantages of our algorithms over other state-of-the-art methods. ER -
APA
Bahadori, M.T., Chang, Y., Long, B. & Liu, Y.. (2014). Scalable Heterogeneous Transfer Ranking. Proceedings of the 3rd International Workshop on Big Data, Streams and Heterogeneous Source Mining: Algorithms, Systems, Programming Models and Applications, in Proceedings of Machine Learning Research 36:214-228 Available from https://proceedings.mlr.press/v36/bahadori14.html.

Related Material