About: This package is an implementation of a linear svm solver with a wide class of regularizations on the svm weight vector (l1, l2, mixed norm l1-lq, adaptive lasso). We provide solvers for the classical single task svm problem and for multi-task with joint feature selection or similarity promoting term.

Initial Announcement on mloss.org.

About: DAL is an efficient and flexibible MATLAB toolbox for sparse learning/reconstruction based on the augmented Lagrangian method.

• 35% faster group lasso.
• Sparse connectivity inference example added (s_test_hsgl.m).
• Non-negative lasso (thanks to Shigeyuki Oba).
• Uses Mark Tygert's pca.m for SVD (PROPACK is not required anymore).

About: A stochastic variant of the mirror descent algorithm employing Langford and Zhang's truncated gradient idea to minimize L1 regularized loss minimization problems for classification and regression.

Fixed major bug in implementation. The components of the iterate where the current example vector is zero were not being updated correctly. Thanks to Jonathan Chang for pointing out the error to us.

About: A (randomized) coordinate descent procedure to minimize L1 regularized loss for classification and regression purposes.

Fixed some I/O bugs. Lines that ended with whitespace were not read correctly in the previous version.