Sciweavers

Share
ICML
2007
IEEE

Exponentiated gradient algorithms for log-linear structured prediction

10 years 2 days ago
Exponentiated gradient algorithms for log-linear structured prediction
Conditional log-linear models are a commonly used method for structured prediction. Efficient learning of parameters in these models is therefore an important problem. This paper describes an exponentiated gradient (EG) algorithm for training such models. EG is applied to the convex dual of the maximum likelihood objective; this results in both sequential and parallel update algorithms, where in the sequential algorithm parameters are updated in an online fashion. We provide a convergence proof for both algorithms. Our analysis also simplifies previous results on EG for max-margin models, and leads to a tighter bound on convergence rates. Experiments on a large-scale parsing task show that the proposed algorithm converges much faster than conjugate-gradient and L-BFGS approaches both in terms of optimization objective and test error.
Amir Globerson, Terry Koo, Xavier Carreras, Michae
Added 17 Nov 2009
Updated 17 Nov 2009
Type Conference
Year 2007
Where ICML
Authors Amir Globerson, Terry Koo, Xavier Carreras, Michael Collins
Comments (0)
books