Abstract

Promising approaches to structured learning problems have recently been developed in the maximum margin framework. Unfortunately, algorithms that are computationally and memory efficient enough to solve large scale problems have lagged behind. We propose using simple subgradient-based techniques for optimizing a regularized risk formulation of these problems in both online and batch settings, and analyze the theoretical convergence, generalization, and robustness properties of the resulting techniques. These algorithms are are simple, memory efficient, fast to converge, and have small regret in the online setting. We also investigate a novel convex regression formulation of structured learning. Finally, we demonstrate the benefits of the subgradient approach on three structured prediction problems. 1

Citations

487	Gradient-based learning applied to document recognition - LeCun, Bottou, et al. - 1998
315	BMaximum-margin Markov networks - Taskar, Guestrin, et al. - 2003
208	Large margin methods for structured and interdependent output variables - Tsochantaridis, Joachims, et al.
192	Minimization Methods for Non-differentiable Functions - Shor - 1985
120	A comparative study of energy minimization methods for Markov random fields - Szeliski, Zabih, et al. - 2006
83	On the generalization ability of on-line learning algorithms - Cesa-Bianchi, Conconi, et al. - 2004
66	M.: Maximum margin planning - Ratliff, Bagnell, et al. - 2006
65	Discriminative learning of markov random fields for segmentation of 3d range data - Anguelov, Taskar, et al. - 2005
65	Logarithmic regret algorithms for online convex optimization - Hazan, Agarwal, et al. - 2006
37	Boosting structured prediction for imitation learning - Ratliff, Bradley, et al. - 2007
33	Convergence rate of incremental subgradient algorithms,” in Stochastic Optimization: Algorithms and Applications - Nedić, Bertsekas - 2000
22	Structured prediction via the extragradient method - Taskar, Lacoste-julien, et al. - 2005
20	An approximation scheme for stochastic linear programming and its application to stochastic integer programs - Shmoys, Swamy
4	A tutorial on energy-based learning. Predicting Structured Data - LeCun, Chopra, et al. - 2006
2	approximate) subgradient methods for structured prediction - Ratliff, Bagnell, et al.