Domain-Dependent Identification of Multiword Expressions

Lecture Notes in Computer Science, 2016

Because most multiword expressions (MWEs), especially verbal ones, are semantically non-compositional, their automatic identification in running text is a prerequisite for semantically-oriented downstream applications. However, recent developments, driven notably by the PARSEME shared task on automatic identification of verbal MWEs, show that this task is harder than related tasks, despite recent contributions both in multilingual corpus annotation and in computational models. In this paper, we analyse possible reasons for this state of affairs. They lie in the nature of the MWE phenomenon, as well as in its distributional properties. We also offer a comparative analysis of the state-of-the-art systems, which exhibit particularly strong sensitivity to unseen data. On this basis, we claim that, in order to make strong headway in MWE identification, the community should bend its mind into coupling identification of MWEs with their discovery, via syntactic MWE lexicons. Such lexicons need not necessarily achieve a linguistically complete modelling of MWEs' behavior, but they should provide minimal morphosyntactic information to cover some potential uses, so as to complement existing MWE-annotated corpora. We define requirements for such a minimal NLP-oriented lexicon, and we propose a roadmap for the MWE community driven by these requirements.

Abstract. We present a recognition algorithm for enumerations of Noun Phrase (NPEs) whose objective is to detect and extract multiword expression (MWE). The algorithm used syntactic rules elaboration from linguistic information aiming to recognize NPEs. This information corresponds to morphological categories (noun, adjective, female, male, etc.). The evaluation takes into account only bigrams found in two different domain corpora of medicine and legal texts.

2008

This paper describes a dataset which provides the platform for a task on the extraction of English verb particle constructions with basic valence information. 5. References Timothy Baldwin. 2005a. The deep lexical acquisition of English verb-particle constructions. Computer Speech and Language, Special Issue on Multiword Expressions, 19(4):398-414. Timothy Baldwin. 2005b. Looking for prepositional verbs in corpus data. In Proc. of the Second ACL-SIGSEM Workshop on the Linguistic Dimensions of Prepositions and their Use in Computational Linguistics Formalisms and Applications, pages 180-9, Colchester, UK. Lou Burnard. 2000. User Reference Guide for the British National Corpus. Technical report, Oxford University Computing Services. Stefan Evert and Brigitte Krenn. 2001. Methods for the qualitative evaluation of lexical association measures. In Proc. of the 39th Annual Meeting of the ACL and 10th Conference of the EACL (ACL-EACL 2001), pages 188-95, Toulouse, France. Dan Flickinger. 2002. On building a more efficient grammar by exploiting types.

A hybrid technique for extraction of bigram and trigram multiword expressions is presented. This technique works in two phases as first statistical technique is applied to filter the extracted bigrams and trigrams from English text, and after it multiword expressions are extracted from this list using some linguistic rules. Two methods for threshold decision in statistical technique are also presented, first is by minimizing the error in classification, and the second is based on maximizing the recall value.

Concepts, Methodologies, Tools, and Applications

This chapter presents a survey of contemporary NLP research on Multiword Expressions (MWEs). MWEs pose a huge problem to precise language processing due to their idiosyncratic nature and diversity of their semantic, lexical, and syntactical properties. The chapter begins by considering MWEs definitions, describes some MWEs classes, indicates problems MWEs generate in language applications and their possible solutions, presents methods of MWE encoding in dictionaries and their automatic detection in corpora. The chapter goes into more detail on a particular MWE class called Verb-Noun Constructions (VNCs). Due to their frequency in corpus and unique characteristics, VNCs present a research problem in their own right. Having outlined several approaches to VNC representation in lexicons, the chapter explains the formalism of Lexical Function as a possible VNC representation. Such representation may serve as a tool for VNCs automatic detection in a corpus. The latter is illustrated on Sp...

Lecture Notes in Computer Science, 2010

Considerable attention has been given to the problem of Multiword Expression (MWE) identification and treatment, for NLP tasks like parsing and generation, to improve the quality of results. Statistical methods have been often employed for MWE identification, as an inexpensive and language independent way of finding co-occurrence patterns. On the other hand, more linguistically motivated methods for identification, which employ information such as POS filters and lexical alignment between languages, can produce more targeted candidate lists. In this paper we propose a hybrid approach that combines the strenghts of different sources of information using a machine learning algorithm to produce more robust and precise results. Automatic evaluation on gold standards shows that the performance of our hybrid method is superior to the individual results of statistical and alignment-based MWE extraction approaches for Portuguese and for English. This method can be used to aid lexicographic work by providing a more targeted MWE candidate list.

Language Resources and Evaluation, 2010

This paper presents the Multiword Expression Toolkit (mwetoolkit), an environment for type and language-independent MWE iden- tification from corpora. The mwetoolkit provides a targeted list of MWE candidates, extracted and filtered according to a number of user-defined criteria and a set of standard statistical association measures. For generating corpus counts, the toolkit provides both a corpus indexation facility and a

Criteria and Gold Standards for the Automatic Identification of Multiword Expressions: Lexicographic Perspective, 2020

The present thesis aims to provide an insight from a lexicographic perspective into the problematic field of research concerned with multiword expressions, criteria according to which they are identified and classified, the identification process thereof itself as well as with the state of the art of gold standards used for their automatic identification. After the introductory part, an explanation as to what lexicography is and what constitutes the lexicographic approach shall follow along with the theoretical background of multiword expressions. An overview of the criteria for the classification and identification of multiword expressions accompanied by the set thereof established by the author of the present thesis is the middle part of the work. Finally, after describing some of the current gold standards for the automatic identification of MWEs, PARSEME’s corpora from three different languages are analyzed and evaluated which leads to the conclusion where a summary of everything done is provided together with certain suggestions for future work on the topic. Keywords: Multiword Expression, Lexicography, NLP, Gold Standard, Automatic Identification, PARSEME

International Journal of Engineering & Technology, 2018

A three phase hybrid method for automatic extraction of English multiword expressions (MWEs) has been proposed. The proposed method is based on linguistic patterns, association and context similarity between constituent words of the MWEs. First, the expressions are extract-ed in the form of N-grams from the raw text and then filtered using well defined linguistic patterns. Next, these expressions are again fil-tered using association score and context similarity score between their constituent words. Two association measures, Dice’s coefficient and PMI have been used for calculating the association score. The context similarity between words has been calculated using Latent Semantic Analysis (LSA) method. The problem of deciding the best value for the cut-off boundary thresholds in statistical methods is quite common. A two phase method of deciding the boundary threshold, using training dataset, has been proposed and employed in the current work. De-tailed performance analysis has b...

Using Neural Networks (NN), Natural Language Processing (NLP) or a dictionary individually to detect noun and verb multiword expressions (MWEs) in an English sentence often results in low accuracy and precision. Our approach combines pre-trained, self-supervised, transformer-based neural networks based on BERT and ALBERT with part-of-speech and sentence dependency, Natural Language Processing algorithms as well as the Unified Compliance, patented dictionary. Experiments were performed on the UCF GRC dataset and publicly available Detecting Minimal Semantic Units and their Meanings (DiM-SUM) dataset . As a result, our approach achieves an F1 score of 73.52% for MWE identification which beats the previous state of the art, models F1 score of 40.76% .