Accelerated training of conditional random fields with stochastic gradient methods

@inproceedings{citeulike:2741191, abstract = {We apply Stochastic Meta-Descent (SMD), a stochastic gradient optimization method with gain vector adaptation, to the training of Conditional Random Fields (CRFs). On several large data sets, the resulting optimizer converges to the same quality of solution over an order of magnitude faster than limited-memory BFGS, the leading method reported to date. We report results for both exact and inexact inference techniques. 1.}, author = {Vishwanathan, Nicol N. }, citeulike-article-id = {2741191}, keywords = {algorithm, machine-learning}, posted-at = {2008-05-02 07:59:14}, priority = {2}, title = {Accelerated training of conditional random fields with stochastic gradient methods}, url = {http://citeseerx.ist.psu.edu/viewdoc/summary;jsessionid=41ED82A909CF37ED4A0E92C1055EE640?cid=3839818}, year = {2006} }

TY - CONF ID - citeulike:2741191 L3 - citeulike-article-id:2741191 N2 - We apply Stochastic Meta-Descent (SMD), a stochastic gradient optimization method with gain vector adaptation, to the training of Conditional Random Fields (CRFs). On several large data sets, the resulting optimizer converges to the same quality of solution over an order of magnitude faster than limited-memory BFGS, the leading method reported to date. We report results for both exact and inexact inference techniques. 1. TI - Accelerated training of conditional random fields with stochastic gradient methods KW - algorithm KW - machine-learning AU - Vishwanathan, Nicol PY - 2006/// UR - http://citeseerx.ist.psu.edu/viewdoc/summary;jsessionid=41ED82A909CF37ED4A0E92C1055EE640?cid=3839818 ER -