Sparse Overlapping Sets Lasso for Multitask Learning and its Application to fMRI Analysis
This addresses feature selection in multitask learning for applications like fMRI analysis, offering a more flexible approach than standard methods, though it is incremental in extending group lasso concepts.
The paper tackles multitask learning where features are organized into overlapping sets and tasks require similar but not identical features, introducing the Sparse Overlapping Sets (SOS) lasso method. Experiments on synthetic and fMRI data show advantages over lasso and group lasso, with derived error bounds and consistency results.
Multitask learning can be effective when features useful in one task are also useful for other tasks, and the group lasso is a standard method for selecting a common subset of features. In this paper, we are interested in a less restrictive form of multitask learning, wherein (1) the available features can be organized into subsets according to a notion of similarity and (2) features useful in one task are similar, but not necessarily identical, to the features best suited for other tasks. The main contribution of this paper is a new procedure called Sparse Overlapping Sets (SOS) lasso, a convex optimization that automatically selects similar features for related learning tasks. Error bounds are derived for SOSlasso and its consistency is established for squared error loss. In particular, SOSlasso is motivated by multi- subject fMRI studies in which functional activity is classified using brain voxels as features. Experiments with real and synthetic data demonstrate the advantages of SOSlasso compared to the lasso and group lasso.