CLARIN-PL Research User-driven Language Technology Infrastructure

Transcription

1 Research User-driven Language Technology Infrastructure Maciej Piasecki Wrocław University of Technology G4.19 Research Group

2 Basic Notions Language Technology (LT) language resources and tools robust in terms of quality and coverage multipurpose component based Language Technology Infrastructure a software framework (architecture or platform) for combining language tools with language resources into processing chains (or pipelines) the defined processing chains are next applied to language data sources interoperability, also with the external systems

3 LT in Humanities and Social Sciences: Barriers Physical language tools and resources are not accessible in Internet Informational descriptions are not available or there is no means for searching Technological lack of commonly accepted standards for LT, lack of a common platform, varieties of technological solutions, insufficient users computers Related to knowledge the use of LT requires programming skills or knowledge from the area of natural language engineering Legal licences for language resources and tools (LRTs) limit their applications

4 CLARIN Support for Humanities & Social Sciences CLARIN is ERIC type consortium of 11 countries (Austria, Bulgaria, Czech Republic, Denmark, Estonia, Germany, Lithuania, The Netherlands, Poland, Portugal, Sweden) and The Dutch Language Union 1 observer: Norway Focus area: Supporting research in Humanities and Social Sciences Users: researchers, PhD students, students and scientific institutions CLARIN Mission To significantly lower the barriers for the use of Language Technology in Humanities & Social Sciences (H&SS) To facilitate or enable research methods based on automated analysis of text and speech resources

5 CLARIN Offer Integration of different LT components into one interoperable system Common, flexible meta-data standard (CMDI) Central searching for resources (Virtual Language Observatory) One sign on and one login into the distributed infrastructure Decreased Physical and Informational Barriers Common standards: promoting, co-ordinating, harmonising Web Services for Language Tools and Resources Decreased Technological Barrier Installation-free, access via Web Applications Decreased Knowledge Barrier Common licences and promotion of the open access Decreased Legal Barrier

6 CLARIN: Portal

7 CLARIN: Virtual Language Observatory

8 CLARIN: Federated Content Search Searching Corpora

9 LTI Development Paradigms Bottom-up a collected offer approach based on linking together the already existing Language Resources and Tools focused on accessibility, technical interoperability and processing chains Top-down following on user-centred design paradigm research applications for H&SS are a starting point Bi-directional linking of Language Resources and Tools combined with the development of research applications

10 Bi-directional LTI Development Idea development of the necessary elements a distributed network infrastructure basic LT processing chain combined with user-centred approach to the development of research applications Top-down part close co-operation with key users from the H&SS domain a metaphor of the Agile-like light weight software designing method with emphasis to prototyping amendments to the shape of the technical basis: LRTs, standards, inspirations, identification of the further user needs, next iterations

11 : the Consortium Polish scientific consortium Wrocław University of Technology, G4.19 Research Group Institute of Computer Science, Polish Academy of Science Polish-Japanese Institute of Information Technology, Chair of Multimedia University of Łódź, PELCRA group at Chair of English Language and Applied Linguistics Institute of Slavic Studies, Polish Academy of Science Wrocław University Goal: implementation of the Polish part of the CLARIN ERIC LTI Follows the bi-directional approach to LTI development

12 : Mission Starting point Several publicly available language resources and tools for Polish, But still many were lacking Deeper technological barrier: restricted applications Pillars: Language Technology Centre the Polish node of the CLARIN distributed infrastructure Complete set of the basic Language Resources & Tools for Polish Research applications for H&SS first set for key users and selected H&SS sub-domains.

13 Language Technology Centre Location in Wrocław University of Technology based on modified D-Space system from Lindat (Czech CLARIN) One sign-on, one login (a member of the Pioneer.id Federation) Advanced repository system for language resources Persistent Identifiers for resources and tools Rich CMDI meta-data CLARIN wide visibility in the central search Interface for Federated Content Search depositing service for researchers from H&SS application for the Data Seal of Approval Adherence to all CLARIN specifications about standards and protocols Web Services for LRTs: the basic processing chain of Polish Prototype system for flexible composition of the natural language processing chains support for developers SOAP & REST interfaces Web Applications for LRTs Knowledge Sharing: expertise and support for the users

14 : Language Resources 1. Polish Morphological Dictionary 2. Polish Speech Corpora 3. Annotated Polish Corpora 4. Bilingual Corpora 5. Polish Historical Corpus 6. Semantic lexicon Wordnet for Polish formal description of lexical meanings 7. Dictionary of Multiword Expressions 8. Bilingual semantic lexicon 9. Lexicon of Proper Names 10.Syntactic-semantic Valency Dictionary 11.Robust syntactic-semantic grammar

15 : Language Resources 1. Polish Morphological Dictionary 2. Polish Speech Corpora 3. Annotated Polish Corpora 4. Bilingual Corpora 5. Polish Historical Corpus 6. Semantic lexicon plwordnet 3.0 formal description of lexical meanings 7. Dictionary of Multiword Expressions 8. Bilingual semantic lexicon 9. Lexicon of Proper Names 10.Syntactic-semantic Valency Dictionary: 11.Robust syntactic-semantic grammar

16 : Language Resources Starting point a set of large resources a huge National Corpus of Polish (1 billion tokens) plwordnet 2.1 a very large wordnet for Polish Korpus Politechniki Wrocławskiej an open Polish corpus with rich annotation Expanded resources plwordnet 3.0 a huge semantic lexicon of Polish a comprehensive description of the Polish lexico-semantic system (~ lemmas, ~ senses) fully mapped to English Princeton WordNet described formally by mapping to an ontology Dictionary of multiword expressions described syntactically NELexicon 2.0 a huge lexicon of Polish Proper Names (2.5 mln)

17 : Language Resources for Polish Expanded resources Conversational corpus (following PELCRA and NKJP) A large semantic valency lexicon for Polish predicative lexical units Newly built resources Transcribed training-testing Polish speech corpus Bi-lingual corpora: Polish-English, Polish-Bulgarian-Russian, Polish-Lithuanian Polish historical corpus (for the years ) Corpora annotated for: meta-data, anaphora, time expressions, spatial expressions, semantic relations and situations

18 plwordnet 2.2 in

19 plwordnet 2.2 in

20 : Language Tools for Polish Systems for searching corpora, especially Polish corpora Spokes for conversational and bilingual corpora Poliqarp 2.0 for richly annotated Historical corpora [New] Text mining (information extraction) Recognition and classification of Proper Names Recognition of anaphoric links Recognition and classification of time expressions and spatial expressions [New] Situation recognition [New] Extraction of multiword expressions (collocations) A generic set of morpho-syntactic tools for Polish that can be adapted to a domain specified by the user [New]

21 : Language Tools for Polish Word Sense Disambiguation based on plwordnet Shallow semantic parser [New] Deep syntactic-semantic parser [New] Tools for the extraction of the semantic-pragmatic information from documents and collections of documents, e.g. keywords [New], semantic relations between text fragments and text summaries

22 Basic Language Tools for Polish 1. Segmentation into tokens and sentences 2. Morphological analysis 3. Morphological guessing of unknown words (both without context and context sensitive) 4. Morpho-syntactic tagging 5. Word Sense Disambiguation 6. Chunker and shallow syntactic parser 7. Named Entity Recognition and disambiguation 8. Co-reference and anaphora resolution 9. Temporal expression recognition 10. Semantic relation recognition 11. Event recognition 12. Shallow semantic parser 13. Deep syntactic parser with disambiguated output: dependency and constituent 14. Deep semantic parser

23 Basic Language Tools for Polish 1. Segmentation into tokens and sentences 2. Morphological analysis 3. Morphological guessing of unknown words (both without context and context sensitive) 4. Morpho-syntactic tagging 5. Word Sense Disambiguation 6. Chunker and shallow syntactic parser 7. Named Entity Recognition and disambiguation 8. Co-reference and anaphora resolution 9. Temporal expression recognition 10. Semantic relation recognition 11. Event recognition 12. Shallow semantic parser 13. Deep syntactic parser with disambiguated output: dependency and constituent 14. Deep semantic parser

24 Basic Language Tools for Polish 1. Segmentation into tokens and sentences 2. Morphological analysis 3. Morphological guessing of unknown words (both without context and context sensitive) 4. Morpho-syntactic tagging 5. Word Sense Disambiguation 6. Chunker and shallow syntactic parser 7. Named Entity Recognition and disambiguation 8. Co-reference and anaphora resolution 9. Temporal expression recognition 10. Semantic relation recognition 11. Event recognition 12. Shallow semantic parser 13. Deep syntactic parser with disambiguated output: dependency and constituent 14. Deep semantic parser

25 : Processing Chain for Polish

26 : Recognition and classification of Proper Names

27 Bi-directional - Top-down Part: First Applications Approaching users already active, interested, working on large textual and speech resources, covering a maximal variety of research areas, e.g. linguistics, literary studies, psychology, political studies and sociology matching the available language tools for Polish the first set of several prototype application illustrating possibilities and facilitating identification of the needs First applications Spokes searching corpora of conversational data A system for collecting Polish text corpora from the Web A open textometric and stylometric system focused on Polish Semantic text classification for sociology Literary Map

28 Spokes (University of Łódź)

29 System for Collecting Polish Text Corpora from the Web Requests from the users revealed gaps in the available technology existing corpus building systems were too sensitive to text encoding errors found in the web not designed for informal corpora like blogs A system for collecting Polish text corpora from the Web had to be constructed: based on tools from the Masaryk University in Brno to detect texts including larger number of errors (by morphological analysis) supports semi-automated extraction of texts from blogs, posts on forums, etc. integrated with tools for processing

30 Open Textometric and Stylometric System System designed for characteristic features of Polish like rich inflection, weakly constrained word order Based on several existing components including Stylo (Eder & Rybicki) Enabling the use of features defined on any level of the linguistic structure: from the level of word forms up to the level of the semantic-pragmatic structures. Available as Web Application and a Web Service Stylometric techniques appear to be applicable in many tasks of H&SS sociology (characteristic features that are for different subgroups), political studies (similarity and differences between political parties), literary studies

31 Semantic Text Classification for Sociology Users: Collegium Civitas, Warsaw Goal Support for large scale analysis of the source materials Automatically annotate documents and text fragments with pre-defined semantic categories Definition of categories by examples Automated semantic grouping of documents and text fragments Support for Corpus building Manual annotation of the learning sub-corpus Automated annotation process Statistical analysis of the results

32 GeTClasS Generalised Text Classification for Sociology

33 Literary Map Users: Digital Humanities Centre of The Institute of Literary Research (Polish Academy of Sciences) Goal Support for using maps in the literary criticism Tool for the identification of all geographical names in the literary text (or a corpus) and mapping them onto a geographical map Tasks 1.Identification and semantic classification of the referring language expressions 2.Disambiguation of the referents 3.Mapping the referents onto a map (geo-location) 4.Recognition of the semantic relations and statistical analysis

34 Literary Map

35 Conclusions Application of LT to the research in Humanities & Social Sciences seem to be much more challenging than in commercial systems! LT for Polish achieved a stage in which valuable support can be provided for research applications Bi-directional approach combines development of the basic, universal set of language tools and resources with inspirations from the research applications

36 Thank you very much for your attention! Supported by the Polish Ministry of Science and Higher Education []

37 Bi-directional: bottom-up part PALC 2014 Łódź LRTs and LRT chains can be useful if the required tools and resources exist, and, they are robust! What is the minimal set of LRTs? What kind of LRTs can be called robust? automated applications in H&SS seem to require high quality of language tools and mostly large coverage of resource BLARK The Basic Language Resource Kit the minimal set of language resources that is necessary to do any precompetitive research and education at all (Krauwer, 2003) and also basic processing chains possible reference point to compare LRTs for different languages