tailieunhanh - Báo cáo khoa học: "Transition-based parsing with Confidence-Weighted Classification"

We show that using confidence-weighted classification in transition-based parsing gives results comparable to using SVMs with faster training and parsing time. We also compare with other online learning algorithms and investigate the effect of pruning features when using confidenceweighted classification. | Transition-based parsing with Confidence-Weighted Classification Martin Haulrich Dept. of International Language Studies and Computational Linguistics Copenhagen Business School Abstract We show that using confidence-weighted classification in transition-based parsing gives results comparable to using SVMs with faster training and parsing time. We also compare with other online learning algorithms and investigate the effect of pruning features when using confidence-weighted classification. 1 Introduction There has been a lot of work on data-driven dependency parsing. The two dominating approaches have been graph-based parsing . MST-parsing McDonald et al. 2005b and transition-based parsing . the MaltParser Nivre et al. 2006a . These two approaches differ radically but have in common that the best results have been obtained using margin-based machine learning approaches. For the MST-parsing MIRA McDonald et al. 2005a McDonald and Pereira 2006 and for transition-based parsing Support-Vector Machines Hall et al. 2006 Nivre et al. 2006b . Dredze et al. 2008 introduce a new approach to margin-based online learning called confidence-weighted classification CW and show that the performance of this approach is comparable to that of Support-Vector Machines. In this work we use confidence-weighted classification with transition-based parsing and show that this leads to results comparable to the state-of-the-art results obtained using SVMs. We also compare training time and the effect of pruning when using confidence-weighted learning. 2 Transition-based parsing Transition-based parsing builds on the idea that parsing can be viewed as a sequence of transitions between states. A transition-based parser deterministic classifier-based parser consists of three essential components Nivre 2008 1. A parsing algorithm 2. A feature model 3. A classifier The focus here is on the classifier but we will briefly describe the parsing algorithm in order to understand .