eVikings%20II

1
eVikings II

• WP3 Supporting RTD in Language Technologies

2
WP3 objectives

• Supporting HLT research in Estonia by development
  of
• human resources
• material resources and research infrastructure
• re-usable language resources
• software

3
WP3 partners

• Laboratory of Phonetics and Speech Technology at
  the Institute of Cybernetics, Tallinn Technical
  University (IOC) WP3 co-ordinator
• Research Group of Computational Linguistics at
  the University of Tartu (UT)
• Language Technology Unit at the Institute of the
  Estonian Language (IEL)
• Laboratory of Acoustics and Audio Signal
  Processing at the Helsinki University of
  Technology (HUT)
• EMT

4
WP3 specific goals

• 1. Enhancement of language resources and
• infrastructure

ELECTRONIC DICTIONARIES
Language resources
LANGUAGE PORTAL
5
WP3 specific goals

• 2. Integration of research groups

6
WP3 specific goals

• 3. Strengthening academic-industrial
  collaboration
• 4. Improving human resources

7
WP3 results

• 1. Language resources
• Morphologically tagged and disambiguated text
  corpus (developed by UT) Deliverable 3.1(1)
• Syntactically tagged text corpus (developed by
  UT) Deliverable 3.1(2)
• Corpus of spoken Estonian and tagged dialogue
  corpus (developed by UT) Deliverable 3.1(3)
• Estonian SpeechDat corpus (developed by IOC)
  Deliverable 3.1(4)
• Language portal Keelevara (developed by IEL)
  Deliverable 3.2
• 19 electronic dictionaries, 2 databases, Handbook
  of the Estonian language, linguistic software
  packages (developed by IEL) Deliverable 3.4

8
WP3 results

• 2. Integration of IOC, UT and IEL
• Sharing of linguistic resources developed by
  partners
• Coordinated research projects funded by the
  national program Estonian language and national
  memory (2004-2008)
• Development of the national program Estonian
  Language Technology (2006-2010)
• Organization of workshops and conferences
• UT doctoral school Linguistics and Language
  Technology (2005-2008)
• Compiling of the Estonian HLT Roadmap
  Deliverable 3.3

9
WP3 results

• 3. Academic-industrial collaboration
• IOC ? EMT
• UT ? Filosoft
• IEL ? Keelevara
• Estonian Language Technology Competence Centre
  (ELTCC) involving
• IOC, UT, IEL ? EMT, Voicecom, Tilde

10
WP3 academic-industrial co-operation

• IOC and EMT
• Support for the development of Estonian SpeechDat
  database
• Supplying the technology and services for the
  collection and processing of speech corpora
• Planning of joint development projects
  (implementation of TTS and ASR)
• Partnership in the application of the Estonian
  Language Technology Competence Centre
• UT and Filosoft
• Dictionary of word frequency of Estonian written
  language
• IEL and Keelevara
• Designing and testing the query system for the
  language portal Keelevara,
• Preparation of new dictionaries and other
  language resources

11
WP3 results

• 4. Development of human resources (1)
• Practical placements in foreign research labs
• Kaarel Kaljurand (UT) at the University of Zurich
  Deliverable 3.13(1)
• Tanel Alumäe (IOC) at the Furui Laboratory at the
  Tokyo Institute of Technology Deliverable
  3.13(2)
• Three workshops on language technology
• Baltic Workshop on Estonian Language Technology
  Applications, April 23, 2003, Tallinn, organised
  by IOC and Archimedes Deliverable 3.12(1)
• First Baltic Conference Human Language
  Technologies the Baltic Perspective Riga,
  Latvia, April 2122, 2004, supported by eVikings
  II Deliverable 3.12(2)
• Tutorials Day on HLT, April 6, 2005, Tallinn,
  organised by IOC and and IEL Deliverable
  3.12(3)

12
WP3 results

• 4. Development of human resources (2)
• International Winter Schools organised by WP2
• International Summer Schools
• 3rd Summer School on Language Technology
  Empirical Methods in Natural Language
  Processing, August 9-14, 2004, Tartu, organised
  by UT Deliverable 3.8(1)
• International Summer School Variation in speech
  production and speech perception, August 10-15,
  2005, Palmse, organised by IOC in co-operation
  with NorFA network Deliverable 3.2(2)

13
WP3 results

• 4. Development of human resources (3)
• Second Baltic Conference on Human Language
  Technologies, Tallinn, Estonia, April 4-6, 2005,
  organised by IOC and IEL Deliverable 3.5
• Participation in international summer/winter
  schools, conferences and workshops Deliverables
  3.9-3.10-3.11
• 82 researchers from WP3 partners participated in
  14 international conferences and workshops
• 51 conference presentations given
• 41 conference papers published in conference
  proceedings

14
WP3 results

• 4. Development of human resources (4)
• Doctoral School of Linguistics and Language
  Technology (2005 - 2008)
• coordinated by UT, partners IEL and IOC
• 7 foreign academic partners
• 2 local industrial partners
• 50 Ph.D. students involved
• 24 defenses planned

15
WP3 international cooperation (IOC)

• HUT Acoustics Lab
• Consulting in the design and development of the
  Estonian SpeechDat corpus
• Participation in organising the Second Baltic
  Conference on Human Language Technology
• Research dealing with improving speech database
  access mechanisms, applicable to the Estonian
  SpeechDat database
• Preparation of a project proposal for FP6 IST
  Call 5, lead by the University of Chalmers,
  Sweden and involves universities from several EU
  countries, including HUT and IOC
• Hosting short-term visits of researchers from IOC
• HUT Computer Science Department
• Joint experiments on developing and testing of
  language-specific ASR-models
• NordForsk
• Network Variation in speech production and
  speech perception - VISPP

16
WP3 international cooperation (IEL)

• The Research Institute for the Languages of
  Finland
• Finnish-Estonian Dictionary I, II (90000 entries,
  Tallinn 2003)
• The Norwegian Estonian Society
• The Norwegian-Estonian Estonian-Norwegian
  dictionary (T. Farbregd, S. Kangur, Ü. Viks.
  Norra-eesti eesti-norra sõnaraamat. Tallinn 1998,
  2005)

17
WP3 international cooperation (UT)

• Nordic projects
• The Nordic Graduate School on Language Technology
• Nordic Treebank Network
• PaNoLA Parsing Nordic Languages
• EU-projects
• EuroTermBank
• Parmenides Ontology driven Temporal Text mining
  on organisational data for extracting temporal
  valid knowledge
• REWERSE Network of Excellence on "Reasoning on
  the Web"

18
WP3 final conclusions

• eVikings II has made a valuable contribution to
  the sustainability and growth of HLT-development
  in Estonia
• Improved quality of human and language resources
• Closer academic cooperation at the national level
• Wider contacts on the international level
• Strengthened links to industry