A Shortest-path Method for Arc-factored Semantic Role Labeling

Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 430–435, October 25-29, 2014, Doha, Qatar. c⃝2014 Association for Computational Linguistics A Shortest-path Method for Arc-factored Semantic Role Labeling Xavier Llu´ıs TALP Research Center Universitat Polit`ecnica de Catalunya xlluis@cs.upc.edu Xavier Carreras Xerox Research Centre Europe xavier.carreras@xrce.xerox.com Llu´ıs M`arquez ALT Research Group Qatar Computing Research Institute lmarquez@qf.org.qa Abstract We introduce a Semantic Role Labeling (SRL) parser that finds semantic roles for a predicate together with the syntactic paths linking predicates and arguments. Our main contribution is to formulate SRL in terms of shortest-path inference, on the as- sumption that the SRL model is restricted to arc-factored features of the syntactic paths behind semantic roles. Overall, our method for SRL is a novel way to ex- ploit larger variability in the syntactic re- alizations of predicate-argument relations, moving away from pipeline architectures. Experiments show that our approach im- proves the robustness of the predictions, producing arc-factored models that per- form closely to methods using unrestricted features from the syntax. 1 Introduction Semantic role labeling (SRL) consists of finding the arguments of a predicate and labeling them with semantic roles (Gildea and Jurafsky, 2002; M`arquez et al., 2008). The arguments fill roles that answer questions of the type “who” did “what” to “whom”, “how”, and “why” for a given sentence predicate. Most approaches to SRL are based on a pipeline strategy, first parsing the sentence to obtain a syntactic tree and then identifying and classifying arguments (Gildea and Jurafsky, 2002; Carreras and M`arquez, 2005). SRL methods critically depend on features of the syntactic structure, and consequently parsing mistakes can harm the quality of semantic role predictions (Gildea and Palmer, 2002). To allevi- ate this dependence, previous work has explored k-best parsers (Johansson and Nugues, 2008), combination systems (Surdeanu et al., 2007) or joint syntactic-semantic models (Johansson, 2009; Henderson et al., 2008; Llu´ıs et al., 2013). In this paper we take a different approach. In our scenario SRL is the end goal, and we as- sume that syntactic parsing is only an intermedi- ate step to extract features to support SRL predic- tions. In this setting we define a model that, given a predicate, identifies each of the semantic roles together with the syntactic path that links the pred- icate with the argument. Thus, following previous work (Moschitti, 2004; Johansson, 2009), we take the syntactic path as the main source of syntac- tic features, but instead of just conditioning on it, we predict it together with the semantic role. The main contribution of this paper is a formulation of SRL parsing in terms of efficient shortest-path in- ference, under the assumption that the SRL model is restricted to arc-factored features of the syntac- tic path linking the argument with the predicate. Our assumption —that features of an SRL model should factor over dependency arcs— is supported by some empirical frequencies. Table 1 shows the most frequent path patterns on CoNLL- 2009 (Hajiˇc et al., 2009) data for several lan- guages, …

Content truncated to 500 words