A Quasi-Newton Approach to Nonsmooth Convex Optimization Problems in Machine Learning
J. Yu, S. Vishwanathan, S.
Günter, and N. N. Schraudolph.
A Quasi-Newton Approach to Nonsmooth Convex Optimization Problems in
Machine Learning. Journal of Machine Learning
Research, 11:1145–1200, 2010.
Short
version
Download
2.5MB | 547.3kB | 1.5MB |
Abstract
We extend the well-known BFGS quasi-Newton method and its memory-limited variant LBFGS to the optimization of nonsmooth convex objectives. This is done in a rigorous fashion by generalizing three components of BFGS to subdifferentials: the local quadratic model, the identification of a descent direction, and the Wolfe line search conditions. We prove that under some technical conditions, the resulting subBFGS algorithm is globally convergent in objective function value. We apply its memory-limited variant (subLBFGS) to $L_2$-regularized risk minimization with the binary hinge loss. To extend our algorithm to the multiclass and multilabel settings, we develop a new, efficient, exact line search algorithm. We prove its worst-case time complexity bounds, and show that it can also be used to extend a recently developed bundle method to the multiclass and multilabel settings. We also apply the direction-finding component of our algorithm to $L_1$-regularized risk minimization with logistic loss. In all these contexts our methods perform comparable to or better than specialized state-of-the-art solvers on a number of publicly available datasets. An open source implementation of our algorithms is freely available.
BibTeX Entry
@article{YuVisGueSch10, author = {Jin Yu and S.~V.~N. Vishwanathan and Simon G\"unter and Nicol N. Schraudolph}, title = {\href{http://nic.schraudolph.org/pubs/YuVisGueSch10.pdf}{ A Quasi-{N}ewton Approach to Nonsmooth Convex Optimization Problems in Machine Learning}}, pages = {1145-1200}, journal = jmlr, volume = 11, year = 2010, b2h_type = {Journal Papers}, b2h_topic = {>Quasi-Newton Methods}, b2h_note = {<a href="b2hd-YuVisGueSch08.html">Short version</a>}, abstract = { We extend the well-known BFGS quasi-Newton method and its memory-limited variant LBFGS to the optimization of nonsmooth convex objectives. This is done in a rigorous fashion by generalizing three components of BFGS to subdifferentials: the local quadratic model, the identification of a descent direction, and the Wolfe line search conditions. We prove that under some technical conditions, the resulting subBFGS algorithm is globally convergent in objective function value. We apply its memory-limited variant (subLBFGS) to $L_{2}$-regularized risk minimization with the binary hinge loss. To extend our algorithm to the multiclass and multilabel settings, we develop a new, efficient, exact line search algorithm. We prove its worst-case time complexity bounds, and show that it can also be used to extend a recently developed bundle method to the multiclass and multilabel settings. We also apply the direction-finding component of our algorithm to $L_{1}$-regularized risk minimization with logistic loss. In all these contexts our methods perform comparable to or better than specialized state-of-the-art solvers on a number of publicly available datasets. An open source implementation of our algorithms is freely available. }}