tailieunhanh - Báo cáo khoa học: "The Best of Both Worlds – A Graph-based Completion Model for Transition-based Parsers"

Transition-based dependency parsers are often forced to make attachment decisions at a point when only partial information about the relevant graph configuration is available. In this paper, we describe a model that takes into account complete structures as they become available to rescore the elements of a beam, combining the advantages of transition-based and graph-based approaches. We also propose an efficient implementation that allows for the use of sophisticated features and show that the completion model leads to a substantial increase in accuracy. We apply the new transition-based parser on typologically different languages such as English, Chinese, Czech, and. | The Best of Both Worlds - A Graph-based Completion Model for Transition-based Parsers Bernd Bohnet and Jonas Kuhn University of Stuttgart Institute for Natural Language Processing bohnet jonas @ Abstract Transition-based dependency parsers are often forced to make attachment decisions at a point when only partial information about the relevant graph configuration is available. In this paper we describe a model that takes into account complete structures as they become available to rescore the elements of a beam combining the advantages of transition-based and graph-based approaches. We also propose an efficient implementation that allows for the use of sophisticated features and show that the completion model leads to a substantial increase in accuracy. We apply the new transition-based parser on typologically different languages such as English Chinese Czech and German and report competitive labeled and unlabeled attachment scores. 1 Introduction Background. A considerable amount of recent research has gone into data-driven dependency parsing and interestingly throughout the continuous process of improvements two classes of parsing algorithms have stayed at the centre of attention the transition-based Nivre 2003 vs. the graph-based approach Eisner 1996 McDonald et al. 2005 .1 The two approaches apply fundamentally different strategies to solve the task of finding the optimal labeled dependency tree over the words of an input sentence where supervised machine learning is used to estimate the scoring parameters on a treebank . The transition-based approach is based on the conceptually and cognitively compelling idea 1More references will be provided in sec. 2. that machine learning . a model of linguistic experience is used in exactly those situations when there is an attachment choice in an otherwise deterministic incremental left-to-right parsing process. As a new word is processed the parser has to decide on one out of a small number of .