An Inequality with Applications to Structured Sparsity and Multitask Dictionary Learning

Andreas Maurer, Massimiliano Pontil, Bernardino Romera-Paredes
; Proceedings of The 27th Conference on Learning Theory, PMLR 35:440-460, 2014.

Abstract

From concentration inequalities for the suprema of Gaussian or Rademacher processes an inequality is derived. It is applied to sharpen existing and to derive novel bounds on the empirical Rademacher complexities of unit balls in various norms appearing in the context of structured sparsity and multitask dictionary learning or matrix factorization. A key role is played by the largest eigenvalue of the data covariance matrix.

Cite this Paper


BibTeX
@InProceedings{pmlr-v35-maurer14, title = {An Inequality with Applications to Structured Sparsity and Multitask Dictionary Learning}, author = {Andreas Maurer and Massimiliano Pontil and Bernardino Romera-Paredes}, booktitle = {Proceedings of The 27th Conference on Learning Theory}, pages = {440--460}, year = {2014}, editor = {Maria Florina Balcan and Vitaly Feldman and Csaba Szepesvári}, volume = {35}, series = {Proceedings of Machine Learning Research}, address = {Barcelona, Spain}, month = {13--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v35/maurer14.pdf}, url = {http://proceedings.mlr.press/v35/maurer14.html}, abstract = {From concentration inequalities for the suprema of Gaussian or Rademacher processes an inequality is derived. It is applied to sharpen existing and to derive novel bounds on the empirical Rademacher complexities of unit balls in various norms appearing in the context of structured sparsity and multitask dictionary learning or matrix factorization. A key role is played by the largest eigenvalue of the data covariance matrix.} }
Endnote
%0 Conference Paper %T An Inequality with Applications to Structured Sparsity and Multitask Dictionary Learning %A Andreas Maurer %A Massimiliano Pontil %A Bernardino Romera-Paredes %B Proceedings of The 27th Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2014 %E Maria Florina Balcan %E Vitaly Feldman %E Csaba Szepesvári %F pmlr-v35-maurer14 %I PMLR %J Proceedings of Machine Learning Research %P 440--460 %U http://proceedings.mlr.press %V 35 %W PMLR %X From concentration inequalities for the suprema of Gaussian or Rademacher processes an inequality is derived. It is applied to sharpen existing and to derive novel bounds on the empirical Rademacher complexities of unit balls in various norms appearing in the context of structured sparsity and multitask dictionary learning or matrix factorization. A key role is played by the largest eigenvalue of the data covariance matrix.
RIS
TY - CPAPER TI - An Inequality with Applications to Structured Sparsity and Multitask Dictionary Learning AU - Andreas Maurer AU - Massimiliano Pontil AU - Bernardino Romera-Paredes BT - Proceedings of The 27th Conference on Learning Theory PY - 2014/05/29 DA - 2014/05/29 ED - Maria Florina Balcan ED - Vitaly Feldman ED - Csaba Szepesvári ID - pmlr-v35-maurer14 PB - PMLR SP - 440 DP - PMLR EP - 460 L1 - http://proceedings.mlr.press/v35/maurer14.pdf UR - http://proceedings.mlr.press/v35/maurer14.html AB - From concentration inequalities for the suprema of Gaussian or Rademacher processes an inequality is derived. It is applied to sharpen existing and to derive novel bounds on the empirical Rademacher complexities of unit balls in various norms appearing in the context of structured sparsity and multitask dictionary learning or matrix factorization. A key role is played by the largest eigenvalue of the data covariance matrix. ER -
APA
Maurer, A., Pontil, M. & Romera-Paredes, B.. (2014). An Inequality with Applications to Structured Sparsity and Multitask Dictionary Learning. Proceedings of The 27th Conference on Learning Theory, in PMLR 35:440-460

Related Material