Semantically Inspired AMR Alignment for the Portuguese Language

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 1595–1600, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 1595 Semantically Inspired AMR Alignment for the Portuguese Language Rafael Torres Anchiˆeta and Thiago Alexandre Salgueiro Pardo Interinstitutional Center for Computational Linguistics (NILC) Institute of Mathematics and Computer Science (ICMC) University of S˜ao Paulo (USP), S˜ao Carlos, SP, Brazil rta@ifpi.edu.br, taspardo@icmc.usp.br Abstract Abstract Meaning Representation (AMR) is a graph-based semantic formalism where the nodes are concepts and edges are relations among them. Most of AMR parsing methods require alignment between the nodes of the graph and the words of the sentence. However, this alignment is not provided by manual an- notations and available automatic aligners fo- cus only on the English language, not perform- ing well for other languages. Aiming to fulfill this gap, we developed an alignment method for the Portuguese language based on a more semantically matched word-concept pair. We performed both intrinsic and extrinsic evalua- tions and showed that our alignment approach outperforms the alignment strategies devel- oped for English, improving AMR parsers, and achieving competitive results with a parser designed for the Portuguese language. 1 Introduction According to Banarescu et al. (2013), Abstract Meaning Representation (AMR) is a semantic meaning representation, which may be encoded as a rooted Direct Acyclic Graph (DAG) where the nodes are concepts, and the edges are relations among them. This representation explicitly details semantics information, as depicted in Figure 1. In this figure, the live-01 node is the root of the graph and city node introduces a named entity. Moreover, :ARGx relations are predicates from the PropBank lexicon (Kingsbury and Palmer, 2002), which encode semantic information according to each PropBank sense. To parse a text into an AMR graph, most of the AMR parsers require alignment between the word (tokens) of the sentence and the nodes of the corresponding graph (see, for instance, (Flanigan et al., 2014; Wang et al., 2015; Zhou et al., 2016; Damonte et al., 2017). However, this anchoring l / live-01 i / i c / city n / name “Barcelona” :ARG0 :location :op1 :name Figure 1: An example of AMR graph for the sentence “I live in Barcelona”. is not provided by manual annotations. Also, the available automatic aligners focus only on the En- glish language (Pourdamghani et al., 2014; Flani- gan et al., 2014; Liu et al., 2018), and they do not perform well for other languages. For Portuguese, for instance, the sentence “N˜ao era surpresa para mim” (It was no surprise to me), the JAMR aligner (Flanigan et al., 2014) produces alignment only be- tween the token surpresa (surprise) and the node surpresa, as shown in Figure 2. # ::snt Não era surpresa para mim # ::alignments 2-3|0 (s / surpresa :polarity - :domain (e / eu)) Figure 2: Alignment produced by JAMR aligner for the sentence “N˜ao era surpresa para mim” (It was no surprise to me). The JAMR aligned only the span 2-3, which is the token surpresa (surprise), with …

Content truncated to 500 words