Robust Structural Metric Learning

Daryl Lim, Gert Lanckriet, Brian McFee
Proceedings of the 30th International Conference on Machine Learning, PMLR 28(1):615-623, 2013.

Abstract

Metric learning algorithms produce a linear transformation of data which is optimized for a prediction task, such as nearest-neighbor classification or ranking. However, when the input data contains a large portion of non-informative features, existing methods fail to identify the relevant features, and performance degrades accordingly. In this paper, we present an efficient and robust structural metric learning algorithm which enforces group sparsity on the learned transformation, while optimizing for structured ranking output prediction. Experiments on synthetic and real datasets demonstrate that the proposed method outperforms previous methods in both high- and low-noise settings.

Cite this Paper

BibTeX
@InProceedings{pmlr-v28-lim13, title = {Robust Structural Metric Learning}, author = {Lim, Daryl and Lanckriet, Gert and McFee, Brian}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {615--623}, year = {2013}, editor = {Dasgupta, Sanjoy and McAllester, David}, volume = {28}, number = {1}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/lim13.pdf}, url = {https://proceedings.mlr.press/v28/lim13.html}, abstract = {Metric learning algorithms produce a linear transformation of data which is optimized for a prediction task, such as nearest-neighbor classification or ranking. However, when the input data contains a large portion of non-informative features, existing methods fail to identify the relevant features, and performance degrades accordingly. In this paper, we present an efficient and robust structural metric learning algorithm which enforces group sparsity on the learned transformation, while optimizing for structured ranking output prediction. Experiments on synthetic and real datasets demonstrate that the proposed method outperforms previous methods in both high- and low-noise settings.} }
Endnote
%0 Conference Paper %T Robust Structural Metric Learning %A Daryl Lim %A Gert Lanckriet %A Brian McFee %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-lim13 %I PMLR %P 615--623 %U https://proceedings.mlr.press/v28/lim13.html %V 28 %N 1 %X Metric learning algorithms produce a linear transformation of data which is optimized for a prediction task, such as nearest-neighbor classification or ranking. However, when the input data contains a large portion of non-informative features, existing methods fail to identify the relevant features, and performance degrades accordingly. In this paper, we present an efficient and robust structural metric learning algorithm which enforces group sparsity on the learned transformation, while optimizing for structured ranking output prediction. Experiments on synthetic and real datasets demonstrate that the proposed method outperforms previous methods in both high- and low-noise settings.
RIS
TY - CPAPER TI - Robust Structural Metric Learning AU - Daryl Lim AU - Gert Lanckriet AU - Brian McFee BT - Proceedings of the 30th International Conference on Machine Learning DA - 2013/02/13 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-lim13 PB - PMLR DP - Proceedings of Machine Learning Research VL - 28 IS - 1 SP - 615 EP - 623 L1 - http://proceedings.mlr.press/v28/lim13.pdf UR - https://proceedings.mlr.press/v28/lim13.html AB - Metric learning algorithms produce a linear transformation of data which is optimized for a prediction task, such as nearest-neighbor classification or ranking. However, when the input data contains a large portion of non-informative features, existing methods fail to identify the relevant features, and performance degrades accordingly. In this paper, we present an efficient and robust structural metric learning algorithm which enforces group sparsity on the learned transformation, while optimizing for structured ranking output prediction. Experiments on synthetic and real datasets demonstrate that the proposed method outperforms previous methods in both high- and low-noise settings. ER -
APA
Lim, D., Lanckriet, G. & McFee, B.. (2013). Robust Structural Metric Learning. Proceedings of the 30th International Conference on Machine Learning, in Proceedings of Machine Learning Research 28(1):615-623 Available from https://proceedings.mlr.press/v28/lim13.html.

Related Material