| { | |
| "paper_id": "1997", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T10:25:31.580907Z" | |
| }, | |
| "title": "KNOWLEDGE BASED MACHINE AIDED TRANSLATION", | |
| "authors": [ | |
| { | |
| "first": "Walther", | |
| "middle": [ | |
| "V" | |
| ], | |
| "last": "Hahn", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Walther", | |
| "middle": [], | |
| "last": "Von Hahn", | |
| "suffix": "", | |
| "affiliation": {}, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "The paper presents and demonstrates the system DB-MAT, a machine aided translation prototype system. Central aim: support of human translation of industrial technical texts by allowing for clarification questions about the domain of the text. Additionally, pictures from the domain are included in retrieval and in the lexicon of the system. Functionality: The translator can select chunks of the source text (or the target text or even from answers to previous queries) and chooses from a nested query menu. The system will derive answers from an internal language independent knowledge base and will present the answer in coherent natural language (at the moment German and Russian). The demo domain is oil/water pollution texts in German and Bulgarian.", | |
| "pdf_parse": { | |
| "paper_id": "1997", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "The paper presents and demonstrates the system DB-MAT, a machine aided translation prototype system. Central aim: support of human translation of industrial technical texts by allowing for clarification questions about the domain of the text. Additionally, pictures from the domain are included in retrieval and in the lexicon of the system. Functionality: The translator can select chunks of the source text (or the target text or even from answers to previous queries) and chooses from a nested query menu. The system will derive answers from an internal language independent knowledge base and will present the answer in coherent natural language (at the moment German and Russian). The demo domain is oil/water pollution texts in German and Bulgarian.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Project Objectives:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "1.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "DBR-MAT (Deutsch-Bulgarisch-Rum\u00e4nisches MAT) is a Machine Aided Translation (MAT) project. Its central aim is to support translators' work by providing domain knowledge, i.e. allowing for clarification questions and integrating pictures.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "1.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The methods applied in this project are an interconnection between the lexicon and a language independent knowledge base of Conceptual Graphs well as a stock of fully indexed pictures.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "1.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The pilot system has been tested in the domain of oil/water pollution and corresponding separation technology.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "1.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "To day translation support mainly means linguistic data and translation memory. If there is any domain knowledge included at all, it is a selection of domain texts. Such explanatory texts may appear as product descriptions by the customer, term definitions in the term bank, or encyclopedic texts on the domain as a whole. These texts, however, are either monolingual or must be translated again to several other ,,standard\" languages to be useful as background knowledge for translators in more than one language pair.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "Even common sense will claim that understanding the domain knowledge included in a text is crucial for the quality of translation. Additionally, questionnaires from translators 4 verified that about 40% of the overall translation time is spent for contents clarifications. The aim of the project is to reduce these 40% by a flexible and user friendly information facility for a translation tool.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "The design principle derived from these facts is to support flexible, language independent, global and integrated access to domain knowledge.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "The central functionality provided by DBR-MAT consists of the following steps:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "The user \u2022 translates in two windows (source and target text),", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "\u2022 selects a piece of text either in the target language or the source language or the previous explanation (text 1,2,3) \u2022 chooses a question type from the \"Information\" menu (e.g., What is?)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "\u2022 receives a natural language answer for the selected text (if it corresponds to a technical term) from an abstract knowledge base.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "\u2022 may ask for corresponding graphics.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Motivation", | |
| "sec_num": "2." | |
| }, | |
| { | |
| "text": "The objects of the KB are primarily concepts such as [DEVICE], but not lexical units, they are secondarily linked to lexical entries. Every object has an internal arbitrary object name (a formal designator) like [OIL_SEPARATOR_1]. Concepts are connected by conceptual relations.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Knowledge Base", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "In the representation language \"Conceptual Graphs\" (Sowal984), concept types and relation types are organized in a type hierarchy.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Knowledge Base", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "An example of two conceptual graphs (contexts) in \"linear notation\": ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Knowledge Base", | |
| "sec_num": "4." | |
| }, | |
| { | |
| "text": "One single coherent domain model (the Knowledge base) supports all meanings of terminological entries in all languages. In contrast to similar approaches the answers to clarification questions are not direct quotes from the knowledge base (in the formal representation), but natural language answers. They are composed from the result of the query by applying graph operations and a natural language generator. The knowledge base (the conceptual structure) is attached to lexicon entries as their \"meaning\".", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "How does the system organize its knowledge?", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "A meaning in DBR-MAT is one (or more) pointer to \"starting nodes\" in the knowledge base (KB). From this starting node the search procedure selects the information specified by the selection from the query menu (What is? exam-pie, characteristics, differences, ...). Due to this technique the borders of word meanings are fuzzy because a user can start querying iteratively out of the system's answers.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "How does the system organize its knowledge?", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "The knowledge base (KB) represents:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "How does the system organize its knowledge?", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "table 2:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "How does the system organize its knowledge?", | |
| "sec_num": "5." | |
| }, | |
| { | |
| "text": "The contents of the KB is acquired from a textbook of the domain rather than from lexical material to make it independent of the test material.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "objects of the knowledge base", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The representation formalism of Conceptual Graphs is well suited for such types of tasks, because it can represent typical properties of language-near objects (terms) and allows for grouping objects (\"contexts\"). This is an interesting method to represent the fact that words in different languages have a different coverage of a conceptual array:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "objects of the knowledge base", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "figure 2: conceptualization in different languages", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "objects of the knowledge base", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "In a given technical domain language A has terms for all objects in this small sample domain, whereas language B and language C depict the concept area with less granularity (or a different conceptualization) each.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "objects of the knowledge base", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "All information sources are interconnected. Lexical items point systematically to conceptual items and to facts in the domain. Figures are accessible from the lexicon as well as from knowledge items. Complex terms (\"Very Large Scale Integration\", e.g.) are handled correctly.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "objects of the knowledge base", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "When answering clarification questions the system follows specific paths in the knowledge base (KB) according to the type of question chosen by the translator. In the following table you find in the first two columns the query types and their subtypes, in the third column the evaluated relations. 2. Find all \"Attr\" relations, 3. Find all \"Char\" relations, 4. Find all \"part_of\" relations. 5. Apply inheritance (if \"More\" is selected under \"Details\", see paragraph 9).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Traversing Rules of the Query Mapper", | |
| "sec_num": "6." | |
| }, | |
| { | |
| "text": "Generating natural language (cf Bontcheva 5 95)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "7.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The result of applying these search rules is already the answer to the query. However, the Conceptual Graphs formalism is not readable by a naive user. Moreover, the result set contains duplicates and trivialities and is not interconnected.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "7.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Therefore a generator (EGEN) evaluates the result set coming from the knowledge base (avoiding duplicates, connecting subparts etc.) and produces a natural language answer. The generation is kept language independent until the last steps (e.g. consulting the KB, traversing the type hierarchy, extracting the relevant knowledge items, determining \"what-to-say\" is language independent). This facilitates the development of new generators of other languages. DBR-MAT presently can run generators for German and Russian.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "7.", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The lexicon of DB-MAT consists of several sublexicons and has the following modular structure:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The DBR-MAT Lexicon", | |
| "sec_num": "8." | |
| }, | |
| { | |
| "text": "{<LexEntry>, <LexMorpho>, <LexSyntax>, <LexText>, <LexTrans>}", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The DBR-MAT Lexicon", | |
| "sec_num": "8." | |
| }, | |
| { | |
| "text": "where LexEntry is the main list of entries: <LexEntry> := lex_entry_<x> ( <Id>, <Entry>, <Type>, <Annotation>, [<CrossRefGroup>, <MorphoGroupId>, <SyntaxGroupId>, SemGroup] ).", | |
| "cite_spans": [ | |
| { | |
| "start": 72, | |
| "end": 172, | |
| "text": "( <Id>, <Entry>, <Type>, <Annotation>, [<CrossRefGroup>, <MorphoGroupId>, <SyntaxGroupId>, SemGroup]", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The DBR-MAT Lexicon", | |
| "sec_num": "8." | |
| }, | |
| { | |
| "text": "<CrossRefGroup> is a set of Ids referring to those lexicon entries which are contained in the entry at hand or which contain this entry. The arguments <MorphoGroupId> and <SyntaxGroupId> refer to the corresponding rule modules of the lexicon, in which their structure is described. <LexSyntax> is only relevant for complex terms and describes their phrase structure.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The DBR-MAT Lexicon", | |
| "sec_num": "8." | |
| }, | |
| { | |
| "text": "In figure 3 a fragment of the lexicon is displayed as it is acquired by the tool Hy-perLAT (see chapter 14). ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The DBR-MAT Lexicon", | |
| "sec_num": "8." | |
| }, | |
| { | |
| "text": "The links from the lexicon entries to the knowledge units are bi-directional:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Linking Lexical Material", | |
| "sec_num": "9." | |
| }, | |
| { | |
| "text": "Lexicon to KB: starting from a lexicon entry (the one selected in the text) the semantics of a term can be calculated by traversing the KB, or KB to lexicon: the generator can search for appropriate terms for KB units The KB can be evaluated to a defined depth depending on the specification of the user (figure 4) less: the local environment of a concept is figure 4. choice of detail consulted only, more: the concept hierarchy is evaluated to include inherited features of more general concept types.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Linking Lexical Material", | |
| "sec_num": "9." | |
| }, | |
| { | |
| "text": "This modular and interconnected representation method has the following advantages:", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Linking Lexical Material", | |
| "sec_num": "9." | |
| }, | |
| { | |
| "text": "\u2022 you see the tacit assumptions included in terms (inherited from the whole terminological system),", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Linking Lexical Material", | |
| "sec_num": "9." | |
| }, | |
| { | |
| "text": "\u2022 you can traverse the whole terminological material in a coherent conceptual system, and \u2022 you can inspect the terminological environment of lexical gaps.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Linking Lexical Material", | |
| "sec_num": "9." | |
| }, | |
| { | |
| "text": "The DBR-MAT paradigm requires the preparation of additional data. The effort to do so, however, must be compared to the effort to prepare and maintain a term bank. It is well known that maintenance of term banks concerning consistency and homogeneity is rather expensive. In DBR-MAT, tools support the terminologist and lexicographer to a high degree, they reduce the amount of work and guarantee consistency and formal correctness.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Costs and benefits for the translator", | |
| "sec_num": "10." | |
| }, | |
| { | |
| "text": "The benefit of the knowledge based approach are", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Costs and benefits for the translator", | |
| "sec_num": "10." | |
| }, | |
| { | |
| "text": "\u2022 new terms can immediately be verbalized in all languages, for which a generator already exists, \u2022 no translation is necessary in term banks, \u2022 term definitions in DBR-MAT are interconnected, \u2022 the user will see not only definitions but further explanations.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Costs and benefits for the translator", | |
| "sec_num": "10." | |
| }, | |
| { | |
| "text": "\u2022 the modularity of components, esp. of the generator, make them reusable for other tasks", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Costs and benefits for the translator", | |
| "sec_num": "10." | |
| }, | |
| { | |
| "text": "The user interface contains two text windows and a variety of menus to access the lexicons and the knowledge base. The menu at the moment contains (besides usual items like File and Edit):", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "DBR-MAT User Interface figure 5: DBR-MAT user interface", | |
| "sec_num": "11." | |
| }, | |
| { | |
| "text": "Note to insert notes and flags to the text, which (in a previous version of DBR-MAT) were included in a translation document, where the translators can see all the flags in context, Information with the facilities to ask questions about the domain, and linguistics, Multilingual for language correspondences.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "DBR-MAT User Interface figure 5: DBR-MAT user interface", | |
| "sec_num": "11." | |
| }, | |
| { | |
| "text": "DBR-MAT is designed to have 5 modes (the vertical columns of figure 6) , three of which are implemented yet. In the center of the graphic the flow of information from queries to the lexicon, the query mapper and the KB in both directions is sketched.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Architecture of DBR-MAT figure 6: System architecture of DBR-MAT", | |
| "sec_num": "12." | |
| }, | |
| { | |
| "text": "In a realistic environment complex systems without additional tools will need as much time for maintenance as they save by their usage. Therefore the maintenance and the acquisition of data must be supported by powerful and easy tools.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "No System without User Tools", | |
| "sec_num": "13." | |
| }, | |
| { | |
| "text": "In DBR-MAT the acquisition and maintenance of the lexicon and the knowledge base can be done (in near future) by a normal user, otherwise the costs for these activities will exceed the benefits of DBR-MAT. The following tools are available or under work (gray) for the languages given in the balloons: This tool relies (except entering the entry itself) only on clicking values from specification tables. This principle rules out any typing errors and allows only reasonable values for a given linguistic item. Endless code lists on the lexicologists table are unnecessary, because everything is displayed on the screen and only when linguistically appropriate. On figure 8 a user is just about defining the encoding table for feminine nouns in German. To make the choice unambiguous the table shows all endings of each classes. ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "No System without User Tools", | |
| "sec_num": "13." | |
| }, | |
| { | |
| "text": "The current laboratory version of DBR-MAT illustrates all important design principles by fully running components (implemented in LPA Prolog). Components in German, Bulgarian, Romanian and Russian are implemented.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Status of Implementation", | |
| "sec_num": "16." | |
| }, | |
| { | |
| "text": "Research is funded by Volkswagen Foundation(1992 -1997) 2 This paper includes scientific results achieved by Dr. Galja Angelova, Kalina Bontcheva (Sofia) and Heike Petermann (Hamburg). See also v.Hahn/Angelova 1996 3 For further information: http: //www.informatik.uni-hamburg.de/Arbeitsbereiche/NATS/projects/db-mat.html", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Kieselbach /Winschiers 1990", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Bontcheva, Kalina: Generation of Multilingual Explanations from Conceptual Graphs. In: Processings of Recent Advances in Natural Language Processing,14-16. Sept. 1995, Tzigov Chark/Bulgaria. S. 184-190.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "DB-MAT: Knowledge Acquisition, Processing and NL Generation using Conceptual Graphs", | |
| "authors": [ | |
| { | |
| "first": "G", | |
| "middle": [ | |
| "/" | |
| ], | |
| "last": "Angelova", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Bontcheva K", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "Proc. ICCS-96", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "ANGELOVA, G. / BONTCHEVA K. (1996): DB-MAT: Knowledge Acquisition, Processing and NL Generation using Conceptual Graphs. To appear in Proc. ICCS-96, August, Lecture Notes of Artificial Intelligence.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Generation of Multilingual Explanations from Conceptual Graphs", | |
| "authors": [ | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Bontcheva", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "Processings of Recent Advances in Natural Language Processing (RANLP)", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "184--190", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "BONTCHEVA, K. (1995): Generation of Multilingual Explanations from Con- ceptual Graphs. In: Processings of Recent Advances in Natural Language Proc- essing (RANLP), 14-16. Sept. 1995, Tzigov Chark/Bulgaria. S. 184-190.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Lexical Ambiguity and the Role of Knowledge Representation in Lexicon Design", | |
| "authors": [ | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Boguraev", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Pustejovsky", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1990, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "BOGURAEV, B./ PUSTEJOVSKY, J. (1990): Lexical Ambiguity and the Role of Knowledge Representation in Lexicon Design. In: COLING-90.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Beyond \"Textbook\" Concept Systems: Handling Multidimensionality in a New Generation of Term Banks", | |
| "authors": [ | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Bowker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "I", | |
| "middle": [], | |
| "last": "Meyer", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "Proceedings of TKE'93: Terminology and Knowledge Engineering", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "BOWKER,L./ MEYER, I. (1993): Beyond \"Textbook\" Concept Systems: Handling Multidimensionality in a New Generation of Term Banks. In: Proceedings of TKE'93: Terminology and Knowledge Engineering.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Translator's Workbench Project. User Requirements Study", | |
| "authors": [ | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Fulford", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Hoege", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Ahmad", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1990, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "FULFORD, H./ HOEGE, M./ AHMAD, K.(1990): Translator's Workbench Proj- ect. User Requirements Study. Report, March.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Innovative Concepts for Machine Aided Translation", | |
| "authors": [ | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Hahn", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "Proceedings VAKKI", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "13--25", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "HAHN, W. (1992): Innovative Concepts for Machine Aided Translation. In: Proceedings VAKKI, Vaasa, Finland, pp. 13-25.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "Combining Terminology, Lexical Semantics and Knowledge Representation in Machine Aided Translation", | |
| "authors": [ | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Hahn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "G", | |
| "middle": [], | |
| "last": "Angelova", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "Proceedings of TKE'96: Terminology and Knowledge Engineering. Frankfurt/M", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "304--314", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "HAHN, W. / ANGELOVA, G.: (1996): Combining Terminology, Lexical Se- mantics and Knowledge Representation in Machine Aided Translation. In Galinski, Chr. and Schmitz, K.-D. (eds.): Proceedings of TKE'96: Terminology and Knowledge Engineering. Frankfurt/M. 304 -314.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Studie zur Anforderungsspezifikation einer computergest\u00fctzten \u00dcbersetzerumgebung. Studienarbeit, Universit\u00e4t Hamburg", | |
| "authors": [ | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Kieselbach", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Winschiers", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1990, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "KIESELBACH, C./WINSCHIERS, H. (1990): Studie zur Anforderungsspezi- fikation einer computergest\u00fctzten \u00dcbersetzerumgebung. Studienarbeit, Uni- versit\u00e4t Hamburg.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "Helping Terminologist Do Knowledge Engineering: Some Linguistic Strategies and Computer Aids", | |
| "authors": [ | |
| { | |
| "first": "I", | |
| "middle": [], | |
| "last": "Meyer", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1994, | |
| "venue": "Actualite Terminologique", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "MEYER, I. (1994): Helping Terminologist Do Knowledge Engineering: Some Linguistic Strategies and Computer Aids. In: Actualite Terminologique, De- cember.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Towards a New Generation of Terminological Resources: An Experiment in Building a Terminological Knowledge Base", | |
| "authors": [ | |
| { | |
| "first": "I", | |
| "middle": [], | |
| "last": "Meyer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Skuce", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Bowker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Eck", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "MEYER, I./ SKUCE, D. / BOWKER,L. /ECK, K. (1992): Towards a New Genera- tion of Terminological Resources: An Experiment in Building a Terminologi- cal Knowledge Base. In: COLING-92.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "CODE4: A Unified System for Managing Conceptual Knowledge", | |
| "authors": [ | |
| { | |
| "first": "D", | |
| "middle": [], | |
| "last": "Skuce", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "T", | |
| "middle": [], | |
| "last": "Lethbridge", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1996, | |
| "venue": "International Journal of Human-Computer Studies and Knowledge Acquisition", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "SKUCE, D. / LETHBRIDGE, T. (1996): CODE4: A Unified System for Managing Conceptual Knowledge. To appear in International Journal of Human- Computer Studies and Knowledge Acquisition.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "Conceptual Structures: Information Processing in Mind and Machine", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Sowa", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1984, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "SOWA, J. (1984): Conceptual Structures: Information Processing in Mind and Machine. Addison Wesley.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "text": "figure 3: DBR-MAT lexicon", | |
| "num": null, | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "FIGREF1": { | |
| "text": "Grammar definition tool of HyperLATOnly a few lines about the lexicon acquisition tool HyperLAT:", | |
| "num": null, | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "FIGREF2": { | |
| "text": "Visual CG browser", | |
| "num": null, | |
| "type_str": "figure", | |
| "uris": null | |
| }, | |
| "TABREF0": { | |
| "text": "", | |
| "num": null, | |
| "html": null, | |
| "type_str": "table", | |
| "content": "<table/>" | |
| }, | |
| "TABREF1": { | |
| "text": "Wellplatte\" is looked up in the lexicon, there the system will find an Id of a KB object, say ,,2245\". Thus the enter point in the KB is defined. Starting from this point the first line of table of the query mapper is executed:", | |
| "num": null, | |
| "html": null, | |
| "type_str": "table", | |
| "content": "<table/>" | |
| } | |
| } | |
| } | |
| } |