1 / 25

OLIF V2

OLIF V2. Gr. Thurmair April 2000. OLIF: Overview. Rationale Principles Entries Descriptions Header Examples Status. text. memory lookup. text. MT. term lookup. text. The Exchange Problem. the same text is processed with different tools how do these tools communicate?.

newton
Télécharger la présentation

OLIF V2

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. OLIF V2 Gr. ThurmairApril 2000

  2. OLIF: Overview • Rationale • Principles • Entries • Descriptions • Header • Examples • Status OLIF April 2000

  3. text memory lookup text MT term lookup text The Exchange Problem • the same text is processed with different tools • how do these tools communicate? Rationale

  4. technical docu MT system 1 user docu MT system 2 marketing docu 1 human / term bank 3 marketing docu 2 human / term bank 4 The Resource Problem Translation Project • how often will the same term be stored? • who pays for redundant maintenance? Rationale

  5. problems use of translation tool 2 The Migration Problem use of translation tool 1 • do you have to re-build your language resources? • who pays for rebuilding lexicons and memories? Rationale

  6. Problems in Exchange • Different purposes of exchange • data import / export • data validation • Different content of exchange • terminological data, different for each system • lexicographical data, different for each system • Different structures of exchange files • trend towards markup structures (SGML/XML) Principles

  7. OLIF: Principles • Keep it simple! • flat feature value structures • standard software environment • Keep it pragmatic! • worry only about what’s there • bottom-up: compare what systems have Principles

  8. OLIF V2 formalism • Define a representation formalism: • XML • general format; processing tools available • but: overhead in markups => not suitable for very large files • Coding standard: UTF-8 • File structure • header: globals, defaults, definitions • body: sequence of entries Principles

  9. OLIF principles • Entries are concept-based • i.e. we describe word senses: different readings -> different entries • Entries have (monolingual) descriptions • both for MT and terminology • Entries have links to other entries • inner-language: crossreference links • intra-language: transfers (multilingual, directed) Principles

  10. Entry Structure • “central” information • definition features • administrative features • monolingual / linguistic feature set • terminological feature set • transfer features • cross-references / links Entries

  11. Definition of the entry • Obligatory features: • language • (only language or also locale?) • canonical form • do we need guidelines (multiwords) • part of speech • only open word classes: N, V, A, Adv, Prep • domain information (semantics) • => a common top level classification? • reading no. Entries

  12. Scope of the descriptions • minimal linguistic descriptions • features which everybody has / needs • minimal terminology descriptions • feature set of e.g. Interval • minimal transfer descriptions • equitype, tests, transfers • minimal thesaurus / ontology relations • ISO standards • additional fields for “personal” use Descriptions

  13. Linguistic Descriptions • Morphological Features • entry type • abbreviation, single word, compound, multiword) • inflection class • enumeration of inflection patterns, per category • gender • (special) number • singulare / plurale tantum • degree / comparative Descriptions

  14. Linguistic descriptions • Syntactic Features • Syntactic Type • (subcategorisation of part-of-speech) • Syntactic Frame • Argument structures (DObj, PObj-for, ...) • (Transitivity) • intransitive, transitive Descriptions

  15. Linguistic Descriptions • Semantic Features • Semantic type • for subclasses only? • who has / needs it? Descriptions

  16. Terminological Descriptions • Minimum needed to validate an entry(Interval) • Definition • Context • Scope • Comment / Note • (validation status) • (a three-level hierarchy) Descriptions

  17. Transfer Descriptions • Equivalence type • full - partial (subset / superset) - none • for reversible entries • Tests and Actions • (to be worked out) • Comment • (Definition of the “target” link) Descriptions

  18. Cross-Reference Descriptions • Linktype • thesaurus relations • broader / narrower / synonym / related • additional customisable relations • abbreviation_for, forbidden, outdated • Definition of the “target” link Descriptions

  19. Administrative Information • Source of Entry • string • Author • creation author • last modofication • Date • creation date • last modification date Descriptions

  20. Header Information • Definition of Encoding • given in the XML statement (UTF-8) • Definition of features / values used • Definition of default values Header

  21. Example (1) <ENTRY> <MONO> <LG> de </LG> <CAN> Brot </CAN> <CAT> noun </CAT> <SA> gv </SA> .... </MONO> </ENTRY> Example

  22. Example 2a <Entry><MONO>.... <CAN> offshore account </CAN> <CAT> noun </CAT> <SA> Money-Laundering </SA> ...</MONO><XFR> <CAN> compte en banque à l‘étranger </CAN> <LG> fr </LG> <CAT> noun </CAT> <SA> Money-Laundering </SA> </XFR><XFR> <CAN> Auslandskonto </CAN> <LG> de </LG> ... </Entry> Example

  23. Example (2b) <Entry><MONO> <CAN> compte en banque à l‘étranger </CAN> <CAT>noun </CAT> <SA>Money-Laundering </SA> ...</MONO><XFR> <CAN>offshore account </CAN> <LG>en </LG> <CAT>noun <SA>Money-Laundering </SA> </XFR><XFR><CAN>Auslandskonto </CAN> <LG>de </LG> ... </Entry> Example

  24. OLIF: Status • Implementation of a central DB • Implementation of OLIF parser & generatorBUT: • different “flavours” of OLIF • dependent on projects (OTELO - Aventinus) • different formalisms (own, SGML, XML) Status

  25. Status • Verification of OLIF • MT lexicons (Logos, T1) • Term Bases (SAPterm, DanTerm) • Converter prototypes • T1 <-> OLIF, Logos <-> OLIF • Term Lookup systems • based on OLIF-type database • Comparisons with other formats Status

More Related