Tractable and Scalable Schatten Quasi-Norm Approximations for Rank Minimization

Fanhua Shang, Yuanyuan Liu, James Cheng
Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, PMLR 51:620-629, 2016.

Abstract

The Schatten quasi-norm was introduced to bridge the gap between the trace norm and rank function. However, existing algorithms are too slow or even impractical for large-scale problems. Motivated by the equivalence relation between the trace norm and its bilinear spectral penalty, we define two tractable Schatten norms, i.e. the bi-trace and tri-trace norms, and prove that they are in essence the Schatten-1/2 and 1/3 quasi-norms, respectively. By applying the two defined Schatten quasi-norms to various rank minimization problems such as MC and RPCA, we only need to solve much smaller factor matrices. We design two efficient linearized alternating minimization algorithms to solve our problems and establish that each bounded sequence generated by our algorithms converges to a critical point. We also provide the restricted strong convexity (RSC) based and MC error bounds for our algorithms. Our experimental results verified both the efficiency and effectiveness of our algorithms compared with the state-of-the-art methods.

Cite this Paper


BibTeX
@InProceedings{pmlr-v51-shang16, title = {Tractable and Scalable Schatten Quasi-Norm Approximations for Rank Minimization}, author = {Shang, Fanhua and Liu, Yuanyuan and Cheng, James}, booktitle = {Proceedings of the 19th International Conference on Artificial Intelligence and Statistics}, pages = {620--629}, year = {2016}, editor = {Gretton, Arthur and Robert, Christian C.}, volume = {51}, series = {Proceedings of Machine Learning Research}, address = {Cadiz, Spain}, month = {09--11 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v51/shang16.pdf}, url = {https://proceedings.mlr.press/v51/shang16.html}, abstract = {The Schatten quasi-norm was introduced to bridge the gap between the trace norm and rank function. However, existing algorithms are too slow or even impractical for large-scale problems. Motivated by the equivalence relation between the trace norm and its bilinear spectral penalty, we define two tractable Schatten norms, i.e. the bi-trace and tri-trace norms, and prove that they are in essence the Schatten-1/2 and 1/3 quasi-norms, respectively. By applying the two defined Schatten quasi-norms to various rank minimization problems such as MC and RPCA, we only need to solve much smaller factor matrices. We design two efficient linearized alternating minimization algorithms to solve our problems and establish that each bounded sequence generated by our algorithms converges to a critical point. We also provide the restricted strong convexity (RSC) based and MC error bounds for our algorithms. Our experimental results verified both the efficiency and effectiveness of our algorithms compared with the state-of-the-art methods.} }
Endnote
%0 Conference Paper %T Tractable and Scalable Schatten Quasi-Norm Approximations for Rank Minimization %A Fanhua Shang %A Yuanyuan Liu %A James Cheng %B Proceedings of the 19th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2016 %E Arthur Gretton %E Christian C. Robert %F pmlr-v51-shang16 %I PMLR %P 620--629 %U https://proceedings.mlr.press/v51/shang16.html %V 51 %X The Schatten quasi-norm was introduced to bridge the gap between the trace norm and rank function. However, existing algorithms are too slow or even impractical for large-scale problems. Motivated by the equivalence relation between the trace norm and its bilinear spectral penalty, we define two tractable Schatten norms, i.e. the bi-trace and tri-trace norms, and prove that they are in essence the Schatten-1/2 and 1/3 quasi-norms, respectively. By applying the two defined Schatten quasi-norms to various rank minimization problems such as MC and RPCA, we only need to solve much smaller factor matrices. We design two efficient linearized alternating minimization algorithms to solve our problems and establish that each bounded sequence generated by our algorithms converges to a critical point. We also provide the restricted strong convexity (RSC) based and MC error bounds for our algorithms. Our experimental results verified both the efficiency and effectiveness of our algorithms compared with the state-of-the-art methods.
RIS
TY - CPAPER TI - Tractable and Scalable Schatten Quasi-Norm Approximations for Rank Minimization AU - Fanhua Shang AU - Yuanyuan Liu AU - James Cheng BT - Proceedings of the 19th International Conference on Artificial Intelligence and Statistics DA - 2016/05/02 ED - Arthur Gretton ED - Christian C. Robert ID - pmlr-v51-shang16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 51 SP - 620 EP - 629 L1 - http://proceedings.mlr.press/v51/shang16.pdf UR - https://proceedings.mlr.press/v51/shang16.html AB - The Schatten quasi-norm was introduced to bridge the gap between the trace norm and rank function. However, existing algorithms are too slow or even impractical for large-scale problems. Motivated by the equivalence relation between the trace norm and its bilinear spectral penalty, we define two tractable Schatten norms, i.e. the bi-trace and tri-trace norms, and prove that they are in essence the Schatten-1/2 and 1/3 quasi-norms, respectively. By applying the two defined Schatten quasi-norms to various rank minimization problems such as MC and RPCA, we only need to solve much smaller factor matrices. We design two efficient linearized alternating minimization algorithms to solve our problems and establish that each bounded sequence generated by our algorithms converges to a critical point. We also provide the restricted strong convexity (RSC) based and MC error bounds for our algorithms. Our experimental results verified both the efficiency and effectiveness of our algorithms compared with the state-of-the-art methods. ER -
APA
Shang, F., Liu, Y. & Cheng, J.. (2016). Tractable and Scalable Schatten Quasi-Norm Approximations for Rank Minimization. Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 51:620-629 Available from https://proceedings.mlr.press/v51/shang16.html.

Related Material