Joint sparsity is widely acknowledged as a powerful structural cue for performing feature selection in setups where variables are expected to demonstrate "grouped" behavior. Such grouped behavior is commonly modeled by Group-Lasso or Multitask Lasso-type problems, where feature selection is effected via ℓ1,q-mixed-norms. Several particular formulations for modeling groupwise sparsity have received substantial attention in the literature; and in some cases, efficient algorithms are also available. Surprisingly, for constrained formulations of fundamental importance (e.g., regression with an ℓ1,∞-norm constraint), highly scalable methods seem to be missing. We address this deficiency by presenting a method based on spectral projected-gradient (SPG) that can tackle ℓ1,q- constrained convex regression problems. The most crucial component of our method is an algorithm for projecting onto ℓ1,q-norm balls. We present several numerical results which show that our methods attain up to 30X speedups on large ℓ1,∞-multitask lasso problems. Even more dramatic are the gains for just the ℓ1,∞-projection subproblem: we observe almost three orders of magnitude speedups compared against the currently standard method. © 2011 Springer-Verlag.
CITATION STYLE
Sra, S. (2011). Fast projections onto ℓ1,q-norm balls for grouped feature selection. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6913 LNAI, pp. 305–317). https://doi.org/10.1007/978-3-642-23808-6_20
Mendeley helps you to discover research relevant for your work.