META-NET and META-SHARE: An Overview
0 likes759 views
Meta-Net is a European initiative aimed at fostering a multilingual information society through language technology and community building. It focuses on preserving linguistic diversity while ensuring equal access to information across Europe, advocating for collaboration among various stakeholders in research and industry. The project includes the development of Meta-Share, a platform for sharing language resources and tools, while promoting innovation in machine translation research.
META-NET and META-SHARE: An Overview
- 1. META-NET and META-SHARE: An Overview Georg Rehm DFKI, Germany georg.rehm@dfki.de Human Language Technologies – The Baltic Perspective Riga, Latvia, October 8, 2010 Co-funded by the 7th Framework Programme of the European Commission through the contract T4ME, grant agreement no.: 249119.
- 2. Outline q q Introduction The three areas of META-NET: (1) Community Building and Development of a SRA (2) META-SHARE (3) Innovative MT Research q The META-NET Website q Extension of META-NET q Collaborations http://www.meta-net.eu 2
- 4. A Socioeconomic Experiment q q q Core component of the European Union: a common market with a single information space. Two dozen national and many regional languages. Europe is an ambitious endeavour and an unprecedented socioeconomic experiment. http://www.meta-net.eu 4
- 5. Two Contradictory Goals? q q u Goal 1: to preserve the diversity of Europe’s cultural heritage (especially its 23 official languages and numerous local variants). Goal 2: to provide all EU-inhabitants with equal access to information and to enable them to participate. Solution: High-performance, robust, crosslingual and multilingual Language Technology! - enables communication across language boundaries, - secures equal access to information, and - supports each language in the advancement of networked LT … … at least in theory. http://www.meta-net.eu 5
- 6. LT: The Current Situation q q Problem: Pace of progress is too slow to bring LT to a level that can fulfil these goals within the next 10 to 20 years. We have to accelerate progress! § But: Language Technology landscape in Europe is highly fragmented. § All stakeholders such as research communities, LT industry, user industries, language communities, funding programmes and policy makers have to team up for a dedicated major push. q We need an alliance of stakeholders dedicated to a large coordinated effort: § META – The Multilingual Europe Technology Alliance § Goals: to assemble and to represent the European LT community, to speak with one voice, to pursue a convincing research agenda. http://www.meta-net.eu 6
- 7. A Simple Mechanism a strong community with ties into other R&D communities with a network into industry and public administration with visions and plans visibility and credibility powerful support from many sectors massive additional resources http://www.meta-net.eu 7
- 8. Examples from other Fields q q q q q ARTEMIS – the European Technology Platform for Advanced Research and Technology for Embedded Intelligence and Systems AAL – The Ambient Assisted Living Joint Programme EPoSS – The European Technology Platform on Smart Systems Integration NEM – The European Technology Platform on Networked and Electronic Media NESSI – The European Technology Platform Networked European Software and Services Initiative http://www.meta-net.eu 8
- 9. META-NET: Initial Funding q T4ME Net: “Technologies for the Multilingual European Information Society” § § § § § § § q Funded under FP 7 13 partners in 10 countries. Coordinated by DFKI. Area: Language-based interaction (ICT-2009.2.2) Grant agreement no.: 249119 Total cost: €7.62 million (EU contribution: €5.99 million) Execution: From Feb 1, 2010 to Jan 31, 2013 (36 months) To avoid multiple brand names, META-NET is used in external communication exclusively. http://www.meta-net.eu 9
- 10. Founding Members q q q q q q q q q q q q q Deutsches Forschungszentrum für Künstliche Intelligenz GmbH, Germany Barcelona Media – Centre d'Innovació, Spain Consiglio Nazionale Ricerche – Instituto di Linguistica Computazionale “Antonio Zampolli”, Italy Institute for Language and Speech Processing, R.C. “Athena”, Greece Charles University in Prague, Czech Republic Centre National de la Recherche Scientifique – Laboratoire d'Informatique pour la Mécanique et les Sci.s de l'Ingénieur, France Universiteit Utrecht, The Netherlands Aalto University, Finland Fondazione Bruno Kessler, Italy Dublin City University, Ireland Rheinisch Westfälische Technische Hochschule Aachen, Germany Jožef Stefan Institute, Slovenia Evaluations and Language Resources Distribution Agency, France http://www.meta-net.eu 10
- 11. META-NET: Objectives q META-NET is a NoE dedicated to fostering the technological foundations of the European multilingual information society: § Build META, a strategic alliance that includes multiple stakeholders to prepare the ground for a large-scale concerted effort. § Strengthen the European research community. § Approach open problems in MT in collaboration with other fields. http://www.meta-net.eu 11
- 12. Community Building Shared Vision and SRA http://www.meta-net.eu 12
- 13. Pillar III: Goals and Actions q q Mobilize researchers, decision makers, users and providers of LT, R&D programmes for cooperation and collaboration in META. Large-scale joint action to Surveying and charting the field Creating a shared vision Preparing a Strategic Research Agenda Building a multiplayer community via external liaisons, promotion, etc. § Identifying stakeholders, informing them about our initiative, dangers and challenges of the situation, LT support for their language etc. § § § § http://www.meta-net.eu 13
- 14. From Visions to the SRA q q q Vision Groups bring together researchers, developers, integrators and (actual or potential, corporate or professional) users of LTbased products, services and applications (ca. 20 members each). A Technology Council will prepare a vision paper “European Multilingual Information Society 2020”, a comprehensive roadmap including individual roadmaps for key socio-economic sectors, lighttower application visions. Finally, a Strategic Research Agenda will be presented by the Council (has to reflect the opinions and needs of all stakeholders). http://www.meta-net.eu 14
- 15. Three Vision Groups q Translation and Localisation (technical documentation, official bulletins, GUI localisation, games, services etc.) § Target stakeholders: large users of translation services, (machine) translation, software companies, game companies, localisation industry q Media and Information Services (audiovisual sector, news, digital libraries, portals, search engines etc.) § Target stakeholders: media industries, search engine providers, archives q Interactive Systems (mobile assistance, dialogue translation, call centres, etc.) § Target stakeholders: mobile software and service providers, telecom industry, call centres http://www.meta-net.eu 15
- 16. Vision Group Meetings q Vision Group Translation and Localisation § 23/07/2010 Berlin, Germany § 28/09/2010 Brussels, Belgium q Vision Group Media and Information Services § 10/09/2010 § 15/10/2010 q Paris, France Barcelona, Spain Vision Group Interactive Systems § 10/09/2010 Paris, France § 05/10/2010 Prague, Czech Republic http://www.meta-net.eu 16
- 17. The Process communication to policy makers funding bodies, public communication in the wider LT community communication and among other stakeholders within META-NET (META-VISION) 2010 http://www.meta-net.eu 2011 2012 17
- 18. Vision Groups: Initial Results q q Initial results (Translation and Localisation Vision Group). Five topics from the first meeting have strong visionary potential: 1. New modes of research studying and involving human translators. 2. A new role for the human in the machine translation process. 3. New computing and deployment models (e.g., specialized services, trust centres, standard data formats, standard platforms such as UIMA). 4. New ways of sharing resources and knowledge in R&D. 5. New, transparent business models. http://www.meta-net.eu 18
- 19. http://www.meta-net.eu/forum q q Please provide input in the discussion forum on our website. Tell us your ideas and priorities! http://www.meta-net.eu 19
- 21. q q q Data has become a key factor in LT (see LREC, corpora mailing list). Expensive data become valuable through sharing. However, the long demanded and well-contemplated instruments for managing and sharing this data are still missing. http://www.meta-net.eu 21
- 22. META-SHARE: Key Features q q q META-SHARE is an open, integrated, secure, and interoperable exchange infrastructure for language data and tools for the Human Language Technologies domain. A marketplace where language data and tools are documented, uploaded and stored in repositories, catalogued and announced, downloaded, exchanged, aiming to support a data economy (includes free and for-a-fee LRs/LTs and also services). Standards-compliant, overcoming format, terminological and semantic differences. http://www.meta-net.eu 22
- 23. META-SHARE Data Centres ELRA, LDC, NICT LT industry, SMEs Regional & national LR projects & initiatives CLARIN http://www.meta-net.eu Acquisition projects PANACEA, TTC, ACCURAT, LET’s MT Academic catalogues & repositories Harvesting initiatives LRE Map, Harvesting Day National data centres 23
- 25. What we’re offering q q q q q q A channel to share and distribute language data and tools. Technical solutions for building your own repositories. Protocols and mechanisms for making the descriptions of your resources (and the actual resources) harvestable. Guidelines and recommendations on standards used in the LR production and documentation processes. Recommendations on data and tools licensing issues. Access to large catalogues of documented, high-quality resources, as well as the actual data and tools. http://www.meta-net.eu 25
- 26. Sneak Peak Sneak Peak (Version 0) http://www.meta-net.eu 26
- 30. Features q Open Source q Single Sign-On q Service-Oriented q Easy Administration q Distributed q Metadata Harvesting q Replication/Backup q Persistent Identifiers (PIDs) q Reporting & Statistics q Intuitive Search http://www.meta-net.eu 30
- 31. META-SHARE: Next Steps q META-SHARE Version 0: November 2010 § First prototype to be demo’ed at META-FORUM 2010 in Brussels (November 17/18). q META-SHARE Version 1: July 2011 § Stable, working version of META-SHARE to be rolled out within the META-NET consortium. q META-SHARE Version 2: February 2012 § Stable version, ready for production use. http://www.meta-net.eu 31
- 33. q Cooperation and building bridges § within the project (machine translation community) § outside (neighbouring research communities such as, for example, machine learning, knowledge management, etc.) q Four scientific goals that have not been addressed yet in a systematic way in the MT community are pursued: § Bringing more semantics into translation § Optimising division of labour in hybrid Machine Translation (consolidation of paradigms) § Exploiting the “context” for translation (e.g., multimedia) § Extending the empirical base for machine translation (using advanced machine learning) http://www.meta-net.eu 33
- 40. META-NET q q T4ME is a funded project. META-NET is a Network of Excellence: § currently based on the initial T4ME consortium. § covers ten countries. q q Multiple European countries are not represented in META-NET. Of course we want to extend the Network of Excellence. http://www.meta-net.eu 40
- 41. Extension of the Network q Upcoming: three new consortia (ICT PSP call). q Objectives (selection): § To collect, polish, upload, share LRs and LTs (via META-SHARE). § To carry out mobilisation and outreach activities in their countries. q q Start: early 2011 META-NET will be extended by all partners of these consortia and also by individual researchers who support the initiative. § Goal: all EU member states and as many of the non-member states as possible to be covered. http://www.meta-net.eu 41
- 42. Extension of META-NET q Current and future composition of META-NET: Year Members Countries 2010 13 10 Current composition. 2011 39 28 Including the three new consortia. 31 Including the three new consortia and a few additional contacts; all EU member states covered. 2011 http://www.meta-net.eu 43 42
- 43. Collaborations Sept. 8, 2010 Communication, Management, Networking, Community Building 43
- 44. CLARIN and META-NET q q q Facilitates research by coordinating and making existing Language Resources and tools available and readily useable for the Social Sciences and Humanities. Offers resources and services to allow computer-aided language processing (e.g., querying data and complex processing of data sets). Focus on eResearch, eScience. Sept. 8, 2010 q q q q q Building and offering results for Language Technology at large. Clear orientation towards ICT (development, innovation and services including commercial). Focus on the distribution of Language Resources (currently). End user: the European citizen. Goal: to address the problem of multilingualism in Europe. Communication, Management, Networking, Community Building 44
- 46. Participate in META! q q q q q Let’s shape the future of Europe’s linguistic and also Language Technology landscape and boost LT support significantly! Researchers, professionals, administrators developing, providing or using LT: please contribute ideas. Educators, journalists, politicians, public administrations and organisations participate in the public discussions on the visions and steps towards a truly multilingual information society. RTD projects can provide and use resources via META-SHARE. Propose projects that contribute to the multilingual digital information space (ICT FP7 Call 7 and SME-DCL; ca. 85 mio. EUR) http://www.meta-net.eu 46
- 47. http://www.meta-forum.eu q q q q q q META-FORUM 2010 – Challenges for Multilingual Europe. November 17/18, Hotel Le Plaza, Brussels, Belgium. Outreach and mobilisation event – primarily aimed at LT industry. Programme includes EC info-session on Call 7 and Call SME-DCL. Ideal networking opportunity. Register soon: interest is very high, we only have limited capacity. http://www.meta-net.eu 47
- 48. Q/A Ļoti pateicos. Tänan väga. Important META-NET Events in 2010 u Language Technology Days, March 22/23, Luxembourg u LREC 2010, May 19-21, Malta u theMETAnk 2010, June 4/5, Berlin u Translingual Europe 2010, June 7, Berlin ICT 2010, Sep 27-29, Brussels: Networking Session “Language Technology for a Multilingual Europe” u Vision Groups (first round of meetings): July 23 (Berlin); Sep 10 (Paris; double feature) u Vision Groups (second round of meetings): Sep 29 (Brussels); Oct 5 (Prague); Oct 15 (Barcelona) u Thank you very much. META-NET Kick-off-Meeting, Feb 22/23, Berlin u Labai ačiū. u META-FORUM 2010 – Challenges for Multilingual Europe, Nov 17/18, Brussels office@meta-net.eu http://www.meta-net.eu http://www.facebook.com/META.Alliance 48
- 50. META-NET and Research Communities semantic cognitive technosystems logies community community http://www.meta-net.eu language technology community machine learning community multimedia content technologies 50
- 51. Community Building policy makers and funding bodies provider industries language communities semantic cognitive technosystems logies community community http://www.meta-net.eu user industries language technology community machine learning community multimedia content technologies 51
- 52. The META Alliance policy makers and funding bodies provider industries language communities semantic cognitive technosystems logies community community http://www.meta-net.eu user industries language technology community machine learning community multimedia content technologies 52