Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/42162
DC FieldValue
dc.titleAn efficient sparse metric learning in high-dimensional space via ℓ1-penalized log-determinant regularization
dc.contributor.authorQi, G.-J.
dc.contributor.authorTang, J.
dc.contributor.authorZha, Z.-J.
dc.contributor.authorChua, T.-S.
dc.contributor.authorZhang, H.-J.
dc.date.accessioned2013-07-04T08:44:55Z
dc.date.available2013-07-04T08:44:55Z
dc.date.issued2009
dc.identifier.citationQi, G.-J.,Tang, J.,Zha, Z.-J.,Chua, T.-S.,Zhang, H.-J. (2009). An efficient sparse metric learning in high-dimensional space via ℓ1-penalized log-determinant regularization. Proceedings of the 26th International Conference On Machine Learning, ICML 2009 : 841-848. ScholarBank@NUS Repository.
dc.identifier.isbn9781605585161
dc.identifier.urihttp://scholarbank.nus.edu.sg/handle/10635/42162
dc.description.abstractThis paper proposes an efficient sparse metric learning algorithm in high dimensional space via an ℓ1-penalized log-determinant regularization. Compare to the most existing distance metric learning algorithms, the proposed algorithm exploits the sparsity nature underlying the intrinsic high dimensional feature space. This sparsity prior of learning distance metric serves to regularize the complexity of the distance model especially in the "less example number p and high dimension d" setting. Theoretically, by analogy to the covariance estimation problem, we find the proposed distance learning algorithm has a consistent result at rate O (√(m2log d)/n) to the target distance matrix with at most m nonzeros per row. Moreover, from the implementation perspective, this ℓ1-penalized log-determinant formulation can be efficiently optimized in a block coordinate descent fashion which is much faster than the standard semi-definite programming which has been widely adopted in many other advanced distance learning algorithms. We compare this algorithm with other state-of-the-art ones on various datasets and competitive results are obtained.
dc.sourceScopus
dc.typeConference Paper
dc.contributor.departmentCOMPUTER SCIENCE
dc.description.sourcetitleProceedings of the 26th International Conference On Machine Learning, ICML 2009
dc.description.page841-848
dc.identifier.isiutNOT_IN_WOS
Appears in Collections:Staff Publications

Show simple item record
Files in This Item:
There are no files associated with this item.

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.