Multi-word Terms Research Articles

Cross-Linguistic Scientific Communication, Contemporary Digital Terminological Neologisms in English, and Translation Issues

The paper develops applied and theoretical results obtained by the author in a long-term, corpus-based research on the cross-linguistic scientific communication and on its most distinctive problem — translating modern scientific texts mainly on digital information processing from English into Russian (and vice versa). These texts are characterized primarily by the growing number of multi-word language-specific new terms. In this respect special attention is paid in the paper to modern authentic «digital» neologisms-clusters in English: to their cross-linguistic idiomatic character, their language-specific types and patterns, to their systematization as well as to their communicative and cognitive distinctions. The paper introduces their equivalent terminological counterparts in Russian and the way they are to be translated into Russian. Some important generalizations and proposals are also introduced concerning new trends in the formation of language-specific «digital» terminological clusters in English and the way they can be rendered into Russian. The paper concludes with outlining some further perspectives in studying cross-linguistic scientific communication, its new terminology and developing its cross-linguistic translation, particularly by digital resources.

Bi-syntax guided transformer network for aspect sentiment triplet extraction

Aspect Sentiment Triplet Extraction is an emerging and challenging task that attempts to present a complete picture of aspect-based sentiment analysis. Prior research efforts mostly leverage various tagging schemes to extract the three elements in a triplet. However, these methods fail to explicitly model the complicated relations between aspects and opinions and the boundaries of multi-word aspects and opinions. In this paper, we propose a bi-syntax guided transformer network in an end-to-end manner to address these challenges. Firstly, we devise three types of representations, including sequence distance representation, constituency distance representation, and dependency distance representation, to learn the comprehensive language representation. Specifically, sequence distance representation utilizes sequence distance between words to enhance the contextual representation. Constituency distance representation adopts constituency distance between words in a constituency tree to capture the intra-span relation between words. Dependency distance representation employs dependency distance between words in a dependency tree to capture the long-distance relation between aspects and opinions. Extensive experiments are conducted on four benchmark datasets to validate the effectiveness of our method. The results demonstrate that the proposed approach achieves better performance than baseline methods. We conduct further detailed analysis to demonstrate that our method effectively handles multi-word terms and overlapping triplets.

Controlled Terms Versus Uncontrolled Terms in Resource Description

The paper comparatively investigates the relation between controlled vocabularies assigned by the expertsin Library of Congress and tags assigned by users in Library Thing database in three subjects, Economics,History and Sociology under Social Science domain. Based on Term matching (S= 14.80 %, E= 12.77 % andH= 8.06 %) and Jaccard similarity coefficient (E= 0.15, S= 0.15 and H= 0.11), we found little matching betweenboth vocabularies. We also found that experts mostly use double-word and multi-word specific topical terms(S= 73.14 %, E= 72.89 % and H= 61.05 %), whereas social taggers mostly use single-word general non-topicalterms (E= 54.88 %, H= 54.21 % and S= 48.55%) and little topical and few personal terms. While comparisonwith LCSH subfield, we found that experts prefer topical terms for all subjects, whereas, taggers only prefer it for Economics and geographic subdivision terms for History and Sociology, but they don’t prefer chronological terms for tagging. Even, experts prefer little title-based terms (H= 196 terms, S= 195 terms and E= 175 terms) but taggers mostly prefer title-based terms (H= 673 terms, S= 564 terms and E= 444 terms) in three subjects. However, the study concludes that both vocabularies are different, but libraries can exploit those uncontrolled vocabularies and can introduce ‘hybrid metadata ecology’ which combines controlled vocabularies, classification and folksonomies for better subject access and retrieval of social science documents.

A syntactic features and interactive learning model for aspect-based sentiment analysis

The aspect-based sentiment analysis (ABSA) consists of two subtasks: aspect term extraction (AE) and aspect term sentiment classification (ASC). Previous research on the AE task has not adequately leveraged syntactic information and has overlooked the issue of multi-word aspect terms in text. Current researchers tend to focus on one of the two subtasks, neglecting the connection between the AE and ASC tasks. Moreover, the problem of error propagation easily occurs between two independent subtasks when performing the complete ABSA task. To address these issues, we present a unified ABSA model based on syntactic features and interactive learning. The proposed model is called syntactic interactive learning based aspect term sentiment classification model (SIASC). To overcome the problem of extracting multi-word aspect terms, the model utilizes part-of-speech features, words features, and dependency features as textual information. Meanwhile, we designs a unified ABSA structure based on the end-to-end framework, reducing the impact of error propagation issues. Interaction learning in the model can establish a connection between the AE task and the ASC task. The information from interactive learning contributes to improving the model’s performance on the ASC task. We conducted an extensive array of experiments on the Laptop14, Restaurant14, and Twitter datasets. The experimental results show that the SIASC model achieved average accuracy of 84.11%, 86.65%, and 78.42% on the AE task, respectively. Acquiring average accuracy of 81.35%, 86.71% and 76.56% on the ASC task, respectively. The SIASC model demonstrates superior performance compared to the baseline model.

Optimized BERT: an effective attention layer based deep learning technique utilizing for multiword term extraction

Optimized BERT: an effective attention layer based deep learning technique utilizing for multiword term extraction

Функционирование фразеологических единиц в текстах судебных решений по уголовным делам (на материале немецкого языка)

This article is dedicated to the study of phraseological units in texts of court decisions in criminal cases within the German legal system. The research objective is to identify legal phraseological units that repeatedly appear in a fixed form in the analyzed legal documents and hold a specialized meaning in terms of professional language. The authors distinguish types of fixed expressions and clichés, organizing them based on their usage in various structural parts of court decisions. The research novelty lies in the first examination of the characteristics of phraseological language tools in the sphere of legal practice based on German court decisions, shedding light on issues related not only to the construction of court decision texts but also to the influence of the content of structural elements on significant text characteristics of legal documents. The analysis revealed a specific set of professionally marked phraseological units regularly present in the examined texts. It was also found that such linguistic clichés act as distinct markers of the compositional organization of court decisions, representing the thematic content of corresponding sections (headings and thematic blocks), predominantly presented either as multi-word terms or as functional verbal constructions and collocations.

Whisper-based spoken term detection systems for search on speech ALBAYZIN evaluation challenge

The vast amount of information stored in audio repositories makes necessary the development of efficient and automatic methods to search on audio content. In that direction, search on speech (SoS) has received much attention in the last decades. To motivate the development of automatic systems, ALBAYZIN evaluations include a search on speech challenge since 2012. This challenge releases several databases that cover different acoustic domains (i.e., spontaneous speech from TV shows, conference talks, parliament sessions, to name a few) aiming to build automatic systems that retrieve a set of terms from those databases. This paper presents a baseline system based on the Whisper automatic speech recognizer for the spoken term detection task in the search on speech challenge held in 2022 within the ALBAYZIN evaluations. This baseline system will be released with this publication and will be given to participants in the upcoming SoS ALBAYZIN evaluation in 2024. Additionally, several analyses based on some term properties (i.e., in-language and foreign terms, and single-word and multi-word terms) are carried out to show the Whisper capability at retrieving terms that convey specific properties. Although the results obtained for some databases are far from being perfect (e.g., for broadcast news domain), this Whisper-based approach has obtained the best results on the challenge databases so far so that it presents a strong baseline system for the upcoming challenge, encouraging participants to improve it.

Word sense disambiguation of acronyms in clinical narratives.

Clinical narratives commonly use acronyms without explicitly defining their long forms. This makes it difficult to automatically interpret their sense as acronyms tend to be highly ambiguous. Supervised learning approaches to their disambiguation in the clinical domain are hindered by issues associated with patient privacy and manual annotation, which limit the size and diversity of training data. In this study, we demonstrate how scientific abstracts can be utilised to overcome these issues by creating a large automatically annotated dataset of artificially simulated global acronyms. A neural network trained on such a dataset achieved the F1-score of 95% on disambiguation of acronym mentions in scientific abstracts. This network was integrated with multi-word term recognition to extract a sense inventory of acronyms from a corpus of clinical narratives on the fly. Acronym sense extraction achieved the F1-score of 74% on a corpus of radiology reports. In clinical practice, the suggested approach can be used to facilitate development of institution-specific inventories.

Isolating Terminology Layers in Complex Linguistic Environments: A Study about Waste Management

Automatic term extraction aims at extracting terminological units from specialized corpora to assist terminographers in developing glossaries, thesauri, and termbases. Unfortunately, traditional methods often overlook the complex relation between terminologies of different subject fields that co-occur in a single specialized corpus. This study illustrates Domain Concept Relatedness, a novel term extraction technique meant to isolate the terminology of a given subject field. We test our technique against the term extraction tool of Sketch Engine and the contrastive approach by applying them to the extraction of waste management terms from a new Italian corpus about waste management legislation. The results show that Domain Concept Relatedness effectively extracts multi-word terms belonging to a given subject field but still fails to extract single-word terms.

TECHNOLOGY FOR AUTOMATED CONSTRUCTION OF DOMAIN DICTIONARIES WITH SPECIAL PROCESSING OF SHORT DOCUMENTS

Context. The task of automating the construction of domain dictionaries in the process of implementing software projects based on the analysis of documents, taking into account their size and presentation form.
 Objective. The goal of the work is to improve the quality of the dictionary based on the use of new technology, including special processing of short documents.
 Method. A model of a short document is proposed, which presents it in the form of three parts: header, content and final. The header and final parts usually contain information not related to the subject area. Therefore, a method for extracting content based on the use of many keywords has been proposed. The size of a short document (its content) does not allow determining the frequency characteristics of words and, therefore, identifying multi-word terms, the share of which reaches 50% of all terms. To make it possible to identify terms in short documents, a method for their clustering is proposed, based on the selection of nouns and the calculation of their frequency characteristics. The resulting clusters are treated as ordinary documents, since their size allows for the selection of multi-word terms. To highlight terms, it is proposed to select sequences of words containing nouns in the text. Analysis of the frequency of repetition of such sequences allows us to identify multi-word terms. To determine the interpretation of terms, a previously developed method of automated search for interpretations in dictionaries was used.
 Results. Based on the proposed model and methods, software was created to build a domain dictionary and a number of experiments were conducted to confirm the effectiveness of the developed solutions.
 Conclusions. The experiments carried out confirmed the performance of the proposed software and allow us to recommend it for use in practice for creating dictionaries of the subject area of various information systems. Prospects for further research may include the construction of corporate search systems based on dictionaries of terms and document clustering.

Trilingual comparative analysis of multi-word terms in karstology: English, French and Croatian

This paper describes the comparative analysis of terms used in karstology in English, French and Croatian. Using the corpus method, co-occurrences of key terms are extracted in order to determine the degree of cross-linguistic equivalence as well as the distribution of certain distinctive features. The results are interpreted with regard to physical-geographical and cultural factors. Unlike a normative terminological description that prescribes exact terminology, this socio-terminological research is primarily descriptive with the aim of understanding the way in which specialized knowledge is structured.

The “New Normal” Terminology

This paper provides a contrastive analysis of COVID-19-related EU legislative texts with emphasis on term variants. The analysis carried out on a parallel corpus consisting of English and Croatian texts has identified multiple examples of variation of predominately multi-word terms at several levels. These include orthographic and grammatical variants, unit shifts and denominative variants creating different meanings. The findings support other contemporary studies into variation within specialized language, highlighting the importance of studying variation to gain understanding about the dynamics of domain knowledge. By focusing on variation in a parallel corpus, conclusions are drawn about the prevalence of variation in a newer EU official language, as opposed to English as the draft text of EU legislation.

Span-based dependency-enhanced graph convolutional network for aspect sentiment triplet extraction

Aspect sentiment triplet extraction task detects three elements of fine-grained sentiment analysis from given sentence, including aspect and opinion terms and their sentiment polarity. Existing methods mainly include tagging-based and span-based methods, where the former show defects on handling overlapped triplets, and the latter could theoretically handle all overlapped triplets but lack of tailored inter-word dependency so that suffer from insufficient span semantic. In this paper, we propose a span-based dependency-enhanced graph convolutional network, which leverages contextual semantic and latent dependency to enrich span representations. Specifically, we devise a latent graph convolutional network to emphasize critical inter-word dependencies and cut off redundant connections in a learnable gating manner, improving the information flow during inter-word interaction. In addition, considering the problem of multi-word term sentiment consistency, we detect effective aspect and opinion terms derived from the output of span enumeration, and introduce term-level interactions by coupling, which meanwhile enables our model to deal with various types of triplets including many-to-one and one-to-many overlapped triplets. Extensive experiments over four benchmark datasets verify that the proposed method outperforms all the baselines with an average F1 improvement of up to 6.13%, and meanwhile shows fine interpretability. The experimental results demonstrate that effectively enhancing token-level and term-level interactions can significantly improve the aspect sentiment triplet extraction performance.

A hybrid part-of-speech tagger with annotated Kurdish corpus: advancements in POS tagging

Abstract With the rapid growth of online content written in the Kurdish language, there is an increasing need to make it machine-readable and processable. Part of speech (POS) tagging is a critical aspect of natural language processing (NLP), playing a significant role in applications such as speech recognition, natural language parsing, information retrieval, and multiword term extraction. This study details the creation of the DASTAN corpus, the first POS-annotated corpus for the Sorani Kurdish dialect. The corpus, containing 74,258 words and thirty-eight tags, employs a hybrid approach utilizing the bigram hidden Markov model in combination with the Kurdish rule-based approach to POS tagging. This approach addresses two key problems that arise with rule-based approaches, namely misclassified words and ambiguity-related unanalyzed words. The proposed approach’s accuracy was assessed by training and testing it on the DASTAN corpus, yielding a 96% accuracy rate. Overall, this study’s findings demonstrate the effectiveness of the proposed hybrid approach and its potential to enhance NLP applications for Sorani Kurdish.

SEWAR: A corpus-based N-gram approach for extracting semantically-related words from Arabic medical corpus

Automatic aggregation of similar words into semantically related groups (or clusters) is of interest to many natural language processing (NLP) applications. Extracting semantically related words and quasi-synonyms from text is a relatively new research area for the under-resourced Arabic language. Previous attempts addressed the problem of single-word term extraction. However, the absence of multiword terms (MWTs) dictionary, ontology, and semantic network makes extracting and identifying high-quality MWTs for the Arabic language a challenging problem. The main goal of this study is to extract corpus-based and coherent MWTs in the form of bigram and trigram sequences as an adequate representation of syntactic and semantic clusters. Therefore, this study contributes to this problem by implementing an algorithm named SEWAR, which uses the FastText algorithm to extract high-quality MWTs from a medical health corpus in Arabic. If put into practice, SEWAR can provide a suitable and helpful solution to classify and direct medical health questions to proper medical practitioners without human intervention. In addition, SEWAR can be applied to plenty of NLP tasks, such as information retrieval, question answering, and text summarization. Three metrics were used to assess the extracted MWTs; the pointwise mutual information (PMI), the cosine similarity, and the clustering purity measure. The results were promising and encouraging to generalize and apply SEWAR to extract MWTs from any Arabic corpus.

MarQO: A query optimizer in multilingual environment for information retrieval in Marathi language

Information retrieval is a crucial component of modern information systems. A significant portion of the vast amount of information stored worldwide is in local languages. While most information retrieval systems are designed primarily for English, there is a growing need for these systems to work with data in languages other than English. Cross Language Information Retrieval (CLIR) systems play a pivotal role in enabling information retrieval across multiple languages. However, these systems often face challenges due to ambiguities in query translation, impacting retrieval accuracy. This paper introduces "MarQO," a query optimizer designed to address these challenges in the context of Marathi language. MarQO employs a multi-stage approach, including lexical processing, extraction of multi-word terms, synonym addition, phrasal translations, utilization of word co-occurrence statistics, and more. By disambiguating query keyword translations, MarQO significantly improves the accuracy of translations, thereby leading to more relevant document retrieval results.

ENGLISH ANATOMICAL TERMS BY ORIGIN IN THE TEXTBOOK HUMAN ANATOMY(I-II)

This research is aimed to identify the particularities of English anatomical terms, based on comparative description of structural forms and in terms of origin. One-word and compound English and Latin anatomical terms are compared. This research was based on the use of anatomical terms in the textbook “Human Anatomy”, published by M. Prives, V. Bushkovich, N. Lisenkov, who offer a progressive view towards? descriptive, evolutionary, functional, and practical anatomy. Medical terms in the textbook can be basically divided into one-word and multi-word terms. One-word terms can be simple words, derived words, compounds, or a combination of derived and compound words. Compound anatomical terms can consist of two-six words. More than a half of the one-word English anatomical terms are formed on the basis of the English lexis. Most English compound anatomical terms are formed using the terms of Latin or Greek languages. Keywords: one-word anatomical terms, compound anatomical terms, borrowed terms, terms formed on the basis of the lexis of one’s own language

Exploring terminological relations between multi-word terms in distributional semantic models

Abstract A term is a lexical unit with specialized meaning in a particular domain. Terms may be simple (STs) or multi-word (MWTs). The organization of terms gives a representation of the structure of domain knowledge, which is based on the relationships between the concepts of the domain. However, relations between MWTs are often underrepresented in terminology resources. This work aims to explore distributional semantic models for capturing terminological relations between multi-word terms through lexical substitution and analogy. The experiments show that the results of the analogy-based method are globally better than those of the one based on lexical substitution and that analogy is well suited to the acquisition of synonymy, antonymy, and hyponymy while lexical substitution performs best for hypernymy.

Machine versus corpus-based translation of multiword terms

Abstract Machine translation (MT) post-editing is an increasingly common practice in the translation industry which is also slowly being applied in the development of terminological resources. However, more studies have been devoted to analyze the practice in a translation scenario than in a terminographic context. Consequently, term-oriented post-editing guidelines are a current need if terminographers are also to become post-editors. With a view to enhancing the multilingual representation of environmental multiword terms (MWTs) in terminological resources, we analyze English–Spanish MWT translation in various generic MT systems. Our aims are: (1) to evaluate MT output in order to check whether it can be of any help to terminographers’ work; (2) to develop an error typology in order to raise terminographers’ awareness; and (3) to use the error typology to sketch a series of basic pre-editing and post-editing rules in a terminographic scenario. A comparison of MT output with the equivalents found in a comparable corpus is also presented. Even though MT often presents errors or unidiomatic choices, it can still serve as a basis for human post-editing, and provided that post-editors are familiarized with the potential errors. Comparable corpora, on the other hand, offer better results, but searches are more time-consuming and equivalents are not always available.

Attention mechanism and skip-gram embedded phrases

This article examines common translation errors that occur in the translation of legal texts. In particular, it focuses on how German texts containing legal terminology are rendered into Modern Greek by the Google translation machine. Our case study is the Google-assisted translation of the original (German) version of the Constitution of the Federal Republic of Germany into Modern Greek. A training method is proposed for phrase extraction based on the occurrence frequency, which goes through the Skip-gram algorithm to be then integrated into the Self Attention Mechanism proposed by Vaswani et al. (2017) in order to minimise human effort and contribute to the development of a robust machine translation system for multi-word legal terms and special phrases. This Neural Machine Translation approach aims at developing vectorised phrases from large corpora and process them for translation. The research direction is to increase the in-domain training data set and enrich the vector dimension with more information for legal concepts (domain specific features).