DOCSLIB.ORG

Nine Terminology Extraction Tools: Are They Useful for Translators?

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Nine Terminology Extraction Tools: Are They Useful for Translators? Nine Terminology Extraction Tools: Are they useful for translators? Hernani Costa, Anna Zaretskaya, Gloria Corpas Pastor, Miriam Seghiri University of Malaga Malaga, Spain fhercos,annazar,g.corpas,[email protected] Abstract such broad range of applications, these tools are often designed for one specific purpose, which Terminology extraction tools have become consequently makes their usage challenging an indispensable resource in education, when employed in a different setting. research and business. Today, users can find a great variety of terminology One of the most important areas where extraction tools of all kinds, and they terminology extraction is extremely helpful is all offer different features. Apart from in the translation industry. Today, more and many other areas, these tools are especially more language service providers (LSP) as well as helpful in the professional translation freelance translators and interpreters understand setting. We do not know, however, if the benefits of automatizing terminology tasks. the existing tools have all the necessary It not only allows them to quickly identify the features for this kind of work. In search for the answer, we make an overview of nine domain of the documents they are dealing with, selected tools available on the market and but also to easily find words and phrases that need find out if they provide the translators’ most to be paid special attention to. While translating favourite features. terminological units, in many cases it is necessary to consider the domain and look up the term equivalents in special resources like terminology 1 Terminology extraction tools and their databases. And in addition, it helps maintain areas of application terminological consistency throughout the project The purpose of terminology extraction tools between all the parts involved: the translator, the (TET) is to help users build terminological LSP and the client. resources in a (semi-)automatic way. The Apart from saving time, another significant need for such resources comes mostly from advantage of using TET instead of manual the growing needs in information management terminology search consists in the possibility to and translation, which make it more and more specify different search criteria, which allows to necessary to have some automated assistance adapt the search query to a particular task. This when performing terminology-related tasks. allows users to see all kinds of information they Companies, freelancers and professionals in need about the term, and also to narrow the search various linguistic fields can resort to these tools and filter the results depending on what they are to, for example build glossaries, thesauri and looking for. As an example, many state-of-the- terminological dictionaries that they use directly art TET offer a possibility to see linguistic and in their work. Moreover, TE is embedded in statistic information about the term, the context a number of natural language processing and where it appears, specify the number of words linguistic research tasks, such as automatic in the term, and many other useful features. indexing, machine translation, information Unfortunately, not every TET offers a full set of extraction, creation of ontologies and knowledge desirable features and settings, which makes it bases, and corpus analysis. Although they have sometimes challenging to find the perfect tool for the task in hand. Apart from the functionalities export glossaries from and to different technology they offer, TET also differ as to the environment environments. In addition, its integration with they work in. For instance, standalone installable SDL Multiterm gives access to many convenient tools require an installation process and work as term-management functions, such as manually independent computer programs. There also exist adding a variety of meta-data information to the web-based tools, which work within a browser. terms, such as synonyms, context, definitions, And finally, there are reusable software that illustrations, part-of-speech tags, URLs, etc., and facilitates the development of larger applications, searching not only the indexed terms but also called frameworks. their descriptive fields. Considering the existing variety, it is not Simple Extractor as its name implies, offers clear how a professional translator is to proceed significantly less functionalities compared to the when choosing a TET suitable for the job. As previous tool. It is a commercial TET developed we will see further, there are some TET that by DAIL Software S.L. for Mac OS, Linux and are specifically created for translators. But do Windows platforms. This clean and easy-to- they have all the necessary characteristics for use standalone Java application was designed to translators? And, furthermore, what exactly are automatically extract the most frequent words these characteristics? and multi-word terms from English, Portuguese, 2 Standalone Terminology Extraction Spanish, French and Russian documents. Simple Tools Extractor not only permits to extract a list of terms (from unigrams up to seven-grams), but Standalone software is probably the most popular also specify the minimum and maximum number type of software today, and TET are no exception. of occurrences of a term. Moreover, Simple Standalone TET are tools that can be installed on Extractor offers an option to load stopword lists, the computer and operate independently of any an advanced search functionality that permits other device or system. to search through the extracted list of terms, to explore all the contexts that a specific term SDL MultiTerm Extract is one of such appears, to edit the term text, to filter the extracted applications. It is a component of SDL terms according to the number of words that form MultiTerm, a commercial terminology them, and to sort the displayed output by any of its management tool that provides one solution fields (frequency, term and context in alphabetical to store, extract and manage multilingual order). Finally, Simple Extractor permits to print terminology. Multiterm exists as a standalone out or export to a file (.pdf, .doc, .csv or .txt) all application, and can also be integrated in SDL the extract terms, as well as their frequencies and Trados Studio. It is one of the few tools that were corresponding contexts. designed specifically to be used by translators and is probably the most well-known TET in TermSuite is an open-source and platform- the translation industry. This TE system locates independent TET written in Java and distributed potential monolingual and bilingual terminology under the Apache License 2.0. It was developed in documents and translation memories using a within the scope of the TTC (Terminology statistic-based method. The user can validate Extraction, Translation Tools and Comparable the extracted candidate terms by looking at a Corpora) project, whose purpose was to design a monolingual or bilingual concordance. A big tool capable of extracting bilingual terminology advantage of this tool is its support for any from comparable corpora in seven languages: language, including Unicode languages. In English, French, German, Spanish, Chinese and addition, it offers a number of functionalities Russian. TermSuite’s architecture is composed by that are useful in different translation scenarios, 3-step modules: the Spotter, the Indexer and the such as ability to compile a dictionary from Aligner. The Spotter module is responsible for parallel texts; flexible filtering that ensures preprocessing the input monolingual corpus, i.e., that only the most frequent candidate terms it performs tokenization, part-of-speech tagging, are extracted; possibility to store an unlimited stemming and lemmatization. Then, the Indexer number of terms in any language; import and module uses both a statistic and a linguistic-based approach to extract monolingual terminology whether to extract only single words (keywords) from a monolingual corpus processed by the or multi-word terminological units (terms). In the Spotter. Finally, the Aligner computes the output, the user can see the keywords or terms, translation of a source terminology into a target links to the five most relevant Wikipedia articles language. The source and target terms required for each of them, the term’s score, its frequency are these already computed by the Indexer in the searched corpus, and its frequency in the module, which means that the previous two reference corpus. There are a variety of search steps should be repeated for the target language. options that can be tuned. For instance, the The user can choose from several alignment user can choose a different reference corpus, options, such as the selection of the maximum decide whether search for words or lemmas, number of translation candidates for a given and accentuate low or high-frequency keywords source term, the use of similarity measures to according to the preferences. The output can compare the contexts of the term in the source be downloaded as a TBX or CSV file. In and the target languages, amongst other advanced order to perform multilingual term extraction the settings. Once all the parameters are set, it is user needs to upload a TMX file with a parallel possible to view and explore all the translation corpus aligned on the sentence or paragraph level. candidates ranked according to their similarity The terminology is first extracted within each score within the tool or use the output XML file language resulting in lists of candidate terms. In for other purposes. the second step, the system searches for such pairs of candidates which co-locate in the parallel 3 Web-Based Terminology Extraction documents most often. The resulting list of Tools candidate pairs (terms in two languages) is then presented to the user. Results can be saved in a Although standalone TET still are predominant TBX or TXT file, which is especially convenient on todays TE applications market, the future for computer-assisted translation tool users. web-based TE technologies will certainly evolve by migrating all standalone features to a web- Translated s.r.l.
Load more

Recommended publications
  • A Comparison of Knowledge Extraction Tools for the Semantic Web
    A Comparison of Knowledge Extraction Tools for the Semantic Web Aldo Gangemi1;2 1 LIPN, Universit´eParis13-CNRS-SorbonneCit´e,France 2 STLab, ISTC-CNR, Rome, Italy. Abstract. In the last years, basic NLP tasks: NER, WSD, relation ex- traction, etc. have been configured for Semantic Web tasks including on- tology learning, linked data population, entity resolution, NL querying to linked data, etc. Some assessment of the state of art of existing Knowl- edge Extraction (KE) tools when applied to the Semantic Web is then desirable. In this paper we describe a landscape analysis of several tools, either conceived specifically for KE on the Semantic Web, or adaptable to it, or even acting as aggregators of extracted data from other tools. Our aim is to assess the currently available capabilities against a rich palette of ontology design constructs, focusing specifically on the actual semantic reusability of KE output. 1 Introduction We present a landscape analysis of the current tools for Knowledge Extraction from text (KE), when applied on the Semantic Web (SW). Knowledge Extraction from text has become a key semantic technology, and has become key to the Semantic Web as well (see. e.g. [31]). Indeed, interest in ontology learning is not new (see e.g. [23], which dates back to 2001, and [10]), and an advanced tool like Text2Onto [11] was set up already in 2005. However, interest in KE was initially limited in the SW community, which preferred to concentrate on manual design of ontologies as a seal of quality. Things started changing after the linked data bootstrapping provided by DB- pedia [22], and the consequent need for substantial population of knowledge bases, schema induction from data, natural language access to structured data, and in general all applications that make joint exploitation of structured and unstructured content.
    [Show full text]
  • A Combined Method for E-Learning Ontology Population Based on NLP and User Activity Analysis
    A Combined Method for E-Learning Ontology Population based on NLP and User Activity Analysis Dmitry Mouromtsev, Fedor Kozlov, Liubov Kovriguina and Olga Parkhimovich ITMO University, St. Petersburg, Russia [email protected], [email protected], [email protected], [email protected] Abstract. The paper describes a combined approach to maintaining an E-Learning ontology in dynamic and changing educational environment. The developed NLP algorithm based on morpho-syntactic patterns is ap- plied for terminology extraction from course tasks that allows to interlink extracted terms with the instances of the system’s ontology whenever some educational materials are changed. These links are used to gather statistics, evaluate quality of lectures’ and tasks’ materials, analyse stu- dents’ answers to the tasks and detect difficult terminology of the course in general (for the teachers) and its understandability in particular (for every student). Keywords: Semantic Web, Linked Learning, terminology extraction, education, educational ontology population 1 Introduction Nowadays reusing online educational resources becomes one of the most promis- ing approaches for e-learning systems development. A good example of using semantics to make education materials reusable and flexible is the SlideWiki system[1]. The key feature of an ontology-based e-learning system is the possi- bility for tutors and students to treat elements of educational content as named objects and named relations between them. These names are understandable both for humans as titles and for the system as types of data. Thus educational materials in the e-learning system thoroughly reflect the structure of education process via relations between courses, modules, lectures, tests and terms.
    [Show full text]
  • Information Extraction Using Natural Language Processing
    INFORMATION EXTRACTION USING NATURAL LANGUAGE PROCESSING Cvetana Krstev University of Belgrade, Faculty of Philology Information Retrieval and/vs. Natural Language Processing So close yet so far Outline of the talk ◦ Views on Information Retrieval (IR) and Natural Language Processing (NLP) ◦ IR and NLP in Serbia ◦ Language Resources (LT) in the core of NLP ◦ at University of Belgrade (4 representative resources) ◦ LR and NLP for Information Retrieval and Information Extraction (IE) ◦ at University of Belgrade (4 representative applications) Wikipedia ◦ Information retrieval ◦ Information retrieval (IR) is the activity of obtaining information resources relevant to an information need from a collection of information resources. Searches can be based on full-text or other content-based indexing. ◦ Natural Language Processing ◦ Natural language processing is a field of computer science, artificial intelligence, and computational linguistics concerned with the interactions between computers and human (natural) languages. As such, NLP is related to the area of human–computer interaction. Many challenges in NLP involve: natural language understanding, enabling computers to derive meaning from human or natural language input; and others involve natural language generation. Experts ◦ Information Retrieval ◦ As an academic field of study, Information Retrieval (IR) is finding material (usually documents) of an unstructured nature (usually text) that satisfies an information need from within large collection (usually stored on computers). ◦ C. D. Manning, P. Raghavan, H. Schutze, “Introduction to Information Retrieval”, Cambridge University Press, 2008 ◦ Natural Language Processing ◦ The term ‘Natural Language Processing’ (NLP) is normally used to describe the function of software or hardware components in computer system which analyze or synthesize spoken or written language.
    [Show full text]
  • A Terminology Extraction System for Technology Related Terms
    TEST: A Terminology Extraction System for Technology Related Terms Murhaf Hossari Soumyabrata Dev John D. Kelleher ADAPT Centre, Trinity College ADAPT Centre, Trinity College ADAPT Centre, Technological Dublin Dublin University Dublin Dublin, Ireland Dublin, Ireland Dublin, Ireland [email protected] [email protected] [email protected] ABSTRACT of applications [6], ranging from e-mail spam filtering, advertising Tracking developments in the highly dynamic data-technology and marketing industry [4, 7], and social media. Particularly, in landscape are vital to keeping up with novel technologies and tools, these realm of online articles and blog articles, the growth in the in the various areas of Artificial Intelligence (AI). However, It is amount of information is almost in an exponential nature. Therefore, difficult to keep track of all the relevant technology keywords. In it is important for the researchers to develop a system that can this paper, we propose a novel system that addresses this problem. automatically parse the millions of documents, and identify the key This tool is used to automatically detect the existence of new tech- technological terms. Such system will greatly reduce the manual nologies and tools in text, and extract terms used to describe these work, and save several man-hours. new technologies. The extracted new terms can be logged as new In this paper, we develop a machine-learning based solution, AI technologies as they are found on-the-fly in the web. It can be that is capable of the discovery and extraction of information re- subsequently classified into the relevant semantic labels and AIdo- lated to AI technologies.
    [Show full text]
  • Ontology Learning and Its Application to Automated Terminology Translation
    Natural Language Processing Ontology Learning and Its Application to Automated Terminology Translation Roberto Navigli and Paola Velardi, Università di Roma La Sapienza Aldo Gangemi, Institute of Cognitive Sciences and Technology lthough the IT community widely acknowledges the usefulness of domain ontolo- A gies, especially in relation to the Semantic Web,1,2 we must overcome several bar- The OntoLearn system riers before they become practical and useful tools. Thus far, only a few specific research for automated ontology environments have ontologies. (The “What Is an Ontology?” sidebar on page 24 provides learning extracts a definition and some background.) Many in the and is part of a more general ontology engineering computational-linguistics research community use architecture.4,5 Here, we describe the system and an relevant domain terms WordNet,3 but large-scale IT applications based on experiment in which we used a machine-learned it require heavy customization. tourism ontology to automatically translate multi- from a corpus of text, Thus, a critical issue is ontology construction— word terms from English to Italian. The method can identifying, defining, and entering concept defini- apply to other domains without manual adaptation. relates them to tions. In large, complex application domains, this task can be lengthy, costly, and controversial, because peo- OntoLearn architecture appropriate concepts in ple can have different points of view about the same Figure 1 shows the elements of the architecture. concept. Two main approaches aid large-scale ontol- Using the Ariosto language processor,6 OntoLearn a general-purpose ogy construction. The first one facilitates manual extracts terminology from a corpus of domain text, ontology engineering by providing natural language such as specialized Web sites and warehouses or doc- ontology, and detects processing tools, including editors, consistency uments exchanged among members of a virtual com- checkers, mediators to support shared decisions, and munity.
    [Show full text]
  • Translate's Localization Guide
    Translate’s Localization Guide Release 0.9.0 Translate Jun 26, 2020 Contents 1 Localisation Guide 1 2 Glossary 191 3 Language Information 195 i ii CHAPTER 1 Localisation Guide The general aim of this document is not to replace other well written works but to draw them together. So for instance the section on projects contains information that should help you get started and point you to the documents that are often hard to find. The section of translation should provide a general enough overview of common mistakes and pitfalls. We have found the localisation community very fragmented and hope that through this document we can bring people together and unify information that is out there but in many many different places. The one section that we feel is unique is the guide to developers – they make assumptions about localisation without fully understanding the implications, we complain but honestly there is not one place that can help give a developer and overview of what is needed from them, we hope that the developer section goes a long way to solving that issue. 1.1 Purpose The purpose of this document is to provide one reference for localisers. You will find lots of information on localising and packaging on the web but not a single resource that can guide you. Most of the information is also domain specific ie it addresses KDE, Mozilla, etc. We hope that this is more general. This document also goes beyond the technical aspects of localisation which seems to be the domain of other lo- calisation documents.
    [Show full text]
  • Generating Domain Terminologies Using Root- and Rule-Based Terms1
    Generating Domain Terminologies using Root- and Rule-Based Terms1 Jacob Collard1, T. N. Bhat2, Eswaran Subrahmanian3,4, Ram D. Sriram3, John T. Elliot2, Ursula R. Kattner2, Carelyn E. Campbell2, Ira Monarch4,5 Independent Consultant, Ithaca, New York1, Materials Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD2, Information Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 3, Carnegie Mellon University, Pittsburgh, PA4, Independent Consultant, Pittsburgh, PA5 Abstract Motivated by the need for flexible, intuitive, reusable, and normalized terminology for guiding search and building ontologies, we present a general approach for generating sets of such terminologies from natural language documents. The terms that this approach generates are root- and rule-based terms, generated by a series of rules designed to be flexible, to evolve, and, perhaps most important, to protect against ambiguity and standardize semantically similar but syntactically distinct phrases to a normal form. This approach combines several linguistic and computational methods that can be automated with the help of training sets to quickly and consistently extract normalized terms. We discuss how this can be extended as natural language technologies improve and how the strategy applies to common use-cases such as search, document entry and archiving, and identifying, tracking, and predicting scientific and technological trends. Keywords dependency parsing; natural language processing; ontology generation; search; terminology generation; unsupervised learning. 1. Introduction 1.1 Terminologies and Semantic Technologies Services and applications on the world-wide web, as well as standards defined by the World Wide Web Consortium (W3C), the primary standards organization for the web, have been integrating semantic technologies into the Internet since 2001 (Koivunen and Miller 2001).
    [Show full text]
  • Terminology Extraction, Translation Tools and Comparable Corpora
    Terminology Extraction, Translation Tools and Comparable Corpora: TTC concept, midterm progress and achieved results Tatiana Gornostay, Anita Gojun, Marion Weller, Ulrich Heid, Emmanuel Morin, Beatrice Daille, Helena Blancafort, Serge Sharoff, Claude Méchoulam To cite this version: Tatiana Gornostay, Anita Gojun, Marion Weller, Ulrich Heid, Emmanuel Morin, et al.. Terminology Extraction, Translation Tools and Comparable Corpora: TTC concept, midterm progress and achieved results. LREC 2012 Workshop on Creating Cross-language Resources for Disconnected Languages and Styles (CREDISLAS), May 2012, Istanbul, Turkey. 4 p. hal-00819909 HAL Id: hal-00819909 https://hal.archives-ouvertes.fr/hal-00819909 Submitted on 9 May 2013 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Terminology Extraction, Translation Tools and Comparable Corpora: TTC concept, midterm progress and achieved results Tatiana Gornostaya, Anita Gojunb, Marion Wellerb, Ulrich Heidb, Emmanuel Morinc, Beatrice Daillec, Helena Blancafortd, Serge Sharoffe, Claude Méchoulamf Tildea, Institute
    [Show full text]
  • Tbxtools: a Free, Fast and Flexible Tool for Automatic Terminology Extraction
    TBXTools: A Free, Fast and Flexible Tool for Automatic Terminology Extraction Antoni Oliver Merce` Vazquez` Universitat Oberta de Catalunya Universitat Oberta de Catalunya [email protected] [email protected] Abstract assisted translation, thesaurus construction, classi- fication, indexing, information retrieval, and also The manual identification of terminology text mining and text summarisation (Heid and Mc- from specialized corpora is a complex task Naught, 1991; Frantzi and Ananiadou, 1996; Vu et that needs to be addressed by flexible al., 2008). tools, in order to facilitate the construction The automatic terminology extraction tools de- of multilingual terminologies which are veloped in recent years allow easier manual term the main resources for computer-assisted extraction from a specialized corpus, which is a translation tools, machine translation or long, tedious and repetitive task that has the risk ontologies. The automatic terminology of being unsystematic and subjective, very costly extraction tools developed so far either use in economic terms and limited by the current a proprietary code or an open source code, available information. However, existing tools that is limited to certain software func- should be improved in order to get more consistent tionalities. To automatically extract terms terminology and greater productivity (Gornostay, from specialized corpora for different pur- 2010). poses such as constructing dictionaries, In the last few years, several term extraction thesauruses or translation memories, we tools have been developed, but most of them are need open source tools to easily integrate language-dependent: French and English –Fastr new functionalities to improve term selec- (Jacquemin, 1999) and Acabit (Daille, 2003); tion. This paper presents TBXTools, a Portuguese –Extracterm (Costa et al., 2004) and free automatic terminology extraction tool ExATOlp (Lopes et al., 2009); Spanish-Basque that implements linguistic and statistical –Elexbi (Hernaiz et al., 2006); Spanish-German methods for multiword term extraction.
    [Show full text]
  • The Interplay Between Lexical Resources and Natural Language Processing
    Tutorial on: The interplay between lexical resources and Natural Language Processing Google Group: goo.gl/JEazYH 1 Luis Espinosa Anke Jose Camacho-Collados Mohammad Taher Pilehvar Google Group: goo.gl/JEazYH 2 Luis Espinosa Anke (linguist) Jose Camacho-Collados (mathematician) Mohammad Taher Pilehvar (computer scientist) 3 www.kahoot.it PIN: 7024700 NAACL 2018 Tutorial: The Interplay between Lexical Resources and Natural Language Processing Camacho-Collados, Espinosa-Anke, Pilehvar 4 QUESTION 1 PIN: 7024700 www.kahoot.it NAACL 2018 Tutorial: The Interplay between Lexical Resources and Natural Language Processing Camacho-Collados, Espinosa-Anke, Pilehvar 5 Outline 1. Introduction 2. Overview of Lexical Resources 3. NLP for Lexical Resources 4. Lexical Resources for NLP 5. Conclusion and Future Directions NAACL 2018 Tutorial: The Interplay between Lexical Resources and Natural Language Processing Camacho-Collados, Espinosa-Anke, Pilehvar 6 INTRODUCTION NAACL 2018 Tutorial: The Interplay between Lexical Resources and Natural Language Processing Camacho-Collados, Espinosa-Anke, Pilehvar 7 Introduction ● “A lexical resource (LR) is a database consisting of one or several dictionaries.” (en.wikipedia.org/wiki/Lexical_resource) ● “What is lexical resource? In a word it is vocabulary and it matters for IELTS writing because …” (dcielts.com/ielts-writing/lexical-resource) ● “The term Language Resource refers to a set of speech or language data and descriptions in machine readable form, used for … ” (elra.info/en/about/what-language-resource )
    [Show full text]
  • Building Ontologies from Folksonomies and Linked Data: Data Structures and Algorithms Technical Report - 22 May 2012
    Building ontologies from folksonomies and linked data: Data structures and Algorithms Technical Report - 22 May 2012 Andrés García-Silva1, Jael García-Castro2, Alexander García3, Oscar Corcho1, Asunción Gómez-Pérez1 1 Ontology Engineering Group, Facultad de Informática, Universidad Politécnica de Madrid, Spain {hgarcia,ocorcho,asun}@fi.upm.es, 2 E-Business & Web Science Research Group, Universitaet der Bundeswehr, Muenchen, Germany [email protected], 3 Biomedical Informatics, Medical Center, University of Arkansas, USA [email protected], Abstract. We present the data structures and algorithms used in the approach for building domain ontologies from folksonomies and linked data. In this approach we extracts domain terms from folksonomies and enrich them with semantic information from the Linked Open Data cloud. As a result, we obtain a domain ontology that combines the emer- gent knowledge of social tagging systems with formal knowledge from Ontologies. 1 Introduction In this report we present the formalization of the data structures and algorithms used in the approach for building ontologies from folksonomies and linked data. In this approach we use folksonomies to gather a domain terminology. First in a term extraction activity we represent folksonomies as a graph which is traversed by using a spreading activation algorithm (see section 2.1). Next in a semantic elicitation activity we identify classes and relations among terms on the terminology relying on linked data sets (see section 2.2). During this activity terms are associated with semantic resources in DBpedia by means of a semantic grounding algorithm. Once terms are grounded to semantic entities we attempt to identify which of those resources correspond to classes in the ontologies that we are using.
    [Show full text]
  • Bilingual Terminology Extraction in Sketch Engine
    Bilingual Terminology Extraction in Sketch Engine Vít Baisa1,2, Barbora Ulipová, Michal Cukr1,2 1 Natural Language Processing Centre, Faculty of Informatics, Masaryk University Botanická 68a, 602 00 Brno, Czech Republic [email protected] 2 Lexical Computing Brighton, United Kingdom and Brno, Czech Republic {vit.baisa,michal.cukr}@sketchengine.co.uk Abstract. We present a method of bilingual terminology extraction from parallel corpora and a few heuristics and experiments with improving the performance of the basic variant of the method. An evaluation is given using a small gold standard manually prepared for English- Czech language pair from DGT translation memory [1]. The bilingual terminology extraction (ABTE3) is available for several languages in Sketch Engine—the corpus management tool [2]. Keywords: terminology extraction, bilingual terminology extraction, Sketch Engine, logDice, parallel corpus 1 Introduction Parallel corpora are valuable resources for machine and computer-assisted translation. Here we explore a possibility of extracting bilingual terminology from parallel corpora combining a monolingual terminology extraction [3] and a co-occurrence statistics [4]. We describe the method and how it is incorporated in the corpus manager tool Sketch Engine. We experimented with parameter tuning and evaluated a few settings using a small gold standard for English- Czech language pair. The following section is a brief survey of topics, methods and tools in ABTE. In Section 3 we describe the basic algorithm for the extraction and in Section 4 how it is integrated in Sketch Engine. In Sections 5 and 6 we evaluate the algorithm and its variants. 2 Related work The monolingual terminology extraction is a well-studied field, and the topic of ABTE has been explored since 90s [5] but recent summarizing publication [6] 3 ATE stands for “automatic terminology extraction”, so we adopt the abbreviation here and add “B” for “bilingual”.
    [Show full text]