Online (and Offline) Robust PCA: Novel Algorithms and Performance Guarantees

Jinchun Zhan, Brian Lois, Han Guo, Namrata Vaswani
Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, PMLR 51:1488-1496, 2016.

Abstract

In this work we develop and study a novel online robust principal components’ analysis (RPCA) algorithm based on the recently introduced ReProCS framework. Our algorithm significantly improves upon the original ReProCS algorithm and it also returns even more accurate offline estimates. The key contribution of this work is a correctness result for this algorithm under relatively mild assumptions. By using extra (but usually valid) assumptions we are able to remove one important limitation of batch RPCA results and two important limitations of a recent result for ReProCS for online RPCA. To the best of our knowledge, this work is among the first correctness results for online RPCA.

Cite this Paper

BibTeX
@InProceedings{pmlr-v51-zhan16, title = {Online (and Offline) Robust PCA: Novel Algorithms and Performance Guarantees}, author = {Zhan, Jinchun and Lois, Brian and Guo, Han and Vaswani, Namrata}, booktitle = {Proceedings of the 19th International Conference on Artificial Intelligence and Statistics}, pages = {1488--1496}, 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/zhan16.pdf}, url = {https://proceedings.mlr.press/v51/zhan16.html}, abstract = {In this work we develop and study a novel online robust principal components’ analysis (RPCA) algorithm based on the recently introduced ReProCS framework. Our algorithm significantly improves upon the original ReProCS algorithm and it also returns even more accurate offline estimates. The key contribution of this work is a correctness result for this algorithm under relatively mild assumptions. By using extra (but usually valid) assumptions we are able to remove one important limitation of batch RPCA results and two important limitations of a recent result for ReProCS for online RPCA. To the best of our knowledge, this work is among the first correctness results for online RPCA.} }
Endnote
%0 Conference Paper %T Online (and Offline) Robust PCA: Novel Algorithms and Performance Guarantees %A Jinchun Zhan %A Brian Lois %A Han Guo %A Namrata Vaswani %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-zhan16 %I PMLR %P 1488--1496 %U https://proceedings.mlr.press/v51/zhan16.html %V 51 %X In this work we develop and study a novel online robust principal components’ analysis (RPCA) algorithm based on the recently introduced ReProCS framework. Our algorithm significantly improves upon the original ReProCS algorithm and it also returns even more accurate offline estimates. The key contribution of this work is a correctness result for this algorithm under relatively mild assumptions. By using extra (but usually valid) assumptions we are able to remove one important limitation of batch RPCA results and two important limitations of a recent result for ReProCS for online RPCA. To the best of our knowledge, this work is among the first correctness results for online RPCA.
RIS
TY - CPAPER TI - Online (and Offline) Robust PCA: Novel Algorithms and Performance Guarantees AU - Jinchun Zhan AU - Brian Lois AU - Han Guo AU - Namrata Vaswani 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-zhan16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 51 SP - 1488 EP - 1496 L1 - http://proceedings.mlr.press/v51/zhan16.pdf UR - https://proceedings.mlr.press/v51/zhan16.html AB - In this work we develop and study a novel online robust principal components’ analysis (RPCA) algorithm based on the recently introduced ReProCS framework. Our algorithm significantly improves upon the original ReProCS algorithm and it also returns even more accurate offline estimates. The key contribution of this work is a correctness result for this algorithm under relatively mild assumptions. By using extra (but usually valid) assumptions we are able to remove one important limitation of batch RPCA results and two important limitations of a recent result for ReProCS for online RPCA. To the best of our knowledge, this work is among the first correctness results for online RPCA. ER -
APA
Zhan, J., Lois, B., Guo, H. & Vaswani, N.. (2016). Online (and Offline) Robust PCA: Novel Algorithms and Performance Guarantees. Proceedings of the 19th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 51:1488-1496 Available from https://proceedings.mlr.press/v51/zhan16.html.

Related Material