M2L at SemEval-2016 Task 8: AMR Parsing with Neural Networks

Proceedings of SemEval-2016, pages 1154–1159, San Diego, California, June 16-17, 2016. c⃝2016 Association for Computational Linguistics M2L at SemEval-2016 Task 8: AMR Parsing with Neural Networks Yevgeniy Puzikov, Daisuke Kawahara, Sadao Kurohashi Graduate School of Informatics, Kyoto University Yoshida-honmachi, Sakyo-ku Kyoto, 606-8501, Japan puzikov@nlp.ist.i.kyoto-u.ac.jp {kuro,dk}@i.kyoto-u.ac.jp Abstract This paper describes our contribution to the SemEval 2016 Workshop. We participated in the Shared Task 8 on Meaning Representa- tion parsing using a transition-based approach, which builds upon the system of Wang et al. (2015a) and Wang et al. (2015b), with addi- tions that utilize a Feedforward Neural Net- work classifier and an enriched feature set. We observed that exploiting Neural Networks in Abstract Meaning Representation parsing is challenging and we could not benefit from it, while the feature enhancements yielded an im- proved performance over the baseline model. 1 Introduction Abstract Meaning Representation (AMR) (Ba- narescu et al., 2013; Dorr et al., 1998) is a semantic formalism which represents sentence meaning in a form of a rooted directed acyclic graph. AMR graph nodes represent concepts, labelled directed edges between the nodes show the relationships between concepts. The AMR formalism was created in or- der to explore the semantics behind natural language units for further analysis and application in various tasks. At the time of writing this paper two AMR parsers are publicly available: graph-based JAMR (Flani- gan et al., 2014) and transition-based CAMR (Wang et al., 2015a; Wang et al., 2015b). The latter has served as our baseline model, which we tried to im- prove by incorporating additional features defined for a wider conditioning context and a neural net- work (NN) classifier. Inspired by the results of Chen and Manning (2014) and Weiss et al. (2015), who obtained state-of-the-art results in transition-based depen- dency parsing using Feedforward Neural Networks (FFNN), and taking into account the transition na- ture of the CAMR model, we performed a series of experiments in the same direction. Neural networks have been successfully applied to many NLP fields and we were curious to examine their potential in the task of AMR parsing. Specifically, we investigated the possibility of constraining the averaged percep- tron algorithm (Collins, 2002) predictions by those of an FFNN at the initial step of the inference pro- cess. 2 Preprocessing We used the Stanford CoreNLP v3.6.0 toolkit (Man- ning et al., 2014) to get named entity (NE) and de- pendency information, the latter in the form of Stan- ford dependencies was obtained from the NN depen- dency parser (Chen and Manning, 2014). We used a publicly available semantic role la- belling (SRL) system with a predicate disambigua- tion module (Bj¨orkelund et al., 2009). The system is a part of MATE tools1, which also include a lem- matizer, a part of speech (POS) tagger, and a de- pendency parser. We use them to obtain lemmas and POS tags. All the tools were used with the pretrained models. Using MALT dependencies instead of or in tan- dem with Stanford dependencies did not change the 1https://storage.googleapis.com/ google-code-archive-downloads/v2/code. google.com/mate-tools …

Content truncated to 500 words