We propose a perceptron-style algorithm for fast discriminative training of structured latent variable model, and analyzed its convergence properties. Our method extends the perceptron algorithm for the learning task with latent dependencies, which may not be captured by traditional models. It relies on Viterbi decoding over latent variables, combined with simple additive updates. Compared to existing probabilistic models of latent variables, our method lowers the training cost significantly yet with comparable or even superior classification accuracy.

Authors

• Xu Sun Peking University
• Takuya Matsuzaki National Institute of Informatics
• Daisuke Okanohara
• Jun’ichi Tsujii

Keywords

No keywords are indexed for this paper.