Riga, Apr HLT in the Baltics, 10 years after 1994 Steven Krauwer ELSNET / Utrecht University (NL)
Riga, Apr Overview About ELSNET Back to 1994 The 2004 picture Where to go from here Fragmentation Your language(s) Europe Human and language resources Concluding remarks
Riga, Apr What is ELSNET European Network in Human Language Technologies (ca 145 academic and industrial member organisations) Funded by the European Commission Created in 1991 Objectives –bringing together the language and speech communities –bringing together academia and andustry –facilitating R&D in language and speech technology Participation is free (conditions on website) Info:
Riga, Apr My picture from 1994 A rich variety of activities Very few interconnections, much fragmentation Poor computing and communication facilities Poor access to publications Clear willingness to collaborate (but few funding opportunities) Access to some EU support through CEE oriented programs
Riga, Apr My picture in 2004 A rich variety of activities Fragmented? See next slide Dramatic improvement in computing and communication facilities Better access to (electronic) publications Clear willingness to collaborate, and (in 10 days time) full access to all EU programs
Riga, Apr My fragmentation test For industry: –count partnerships between HLT R&D parties For academia: –count co-authorships of papers across institutions in the proceedings –count bibliographical references to others than (co-)authors (and their spiritual leaders) –homework for tonight!
Riga, Apr Where to go from here? Defragmentation by enhancing local cohesion Watch your language Join the European scene Take care of your resources
Riga, Apr Defragmentation Make collaboration across institutions priority on your agenda Use joint authorships as a measure Analyze your bibliographical references from time to time Exploit local and international programmes to move people around
Riga, Apr Watch your language Your language is a precious possession, but if you don’t take care of it, no one will Speaking English may be a rather common thing for educated people, but for many people it isn’t – use your technologies to avoid a two class society You are the best experts in your own language – exploit this before others do it!
Riga, Apr Join the European scene Lots of EU funding opportunities, but –Language and speech technology as such have disappeared from the FP6 research agenda: they will only be present as embedded technologies in other, much broader projects –the move from ‘spreading excellence’ (in earlier FPs) to ‘boosting excellence (in FP6) has severe consequences Look at the participation figures in the successful projects of the 1 st IST Call for Proposals for Interface Technologies
Riga, Apr st IST Call (Interfaces) Participation per country
Riga, Apr st IST Call (Interfaces) Top 8 countries
Riga, Apr How to react to this? Not recommended: –Accept the defeat of the smaller languages –Hope that the EU will do something special for the smaller languages to compensate Recommended: –Make yourselves stronger –Be patient: language and speech technology may make a glorious come-back in FP7 – if only to prevent the EU from collapsing under the language problem
Riga, Apr How to make yourselves stronger Increase internal cohesion (defragmentation) Exploit the fact that you are the top experts in your own language Exploit the fact that many wheels have already been invented (although they may not yet be round enough for your own language) Be prepared to look across discipline boundaries Team up with stronger parties in Europe by –staff and student exchanges (exploit mobility funds) –making yourselves visible (at conferences and events) –focusing on a niche of expertise
Riga, Apr Take care of your resources Human resources: –Develop HLT curricula, preferably on a very interdisciplinary basis (e.g Euro-Masters) –Develop training programmes for professionals already working in the field Language resources: –Ensure the availability of proper resources for your own language, at least for the purpose of research, precompetitive development and training –Join our BLARK action
Riga, Apr The BLARK Define the minimal set of language resources that is necessary to do any precompetitive R&D and education at all for a language (the Basic Language Resource Kit or BLARK) Determine which components are already available Make a priority plan to complete the BLARK
Riga, Apr What makes the BLARK special? The idea is to make a common generic BLARK definition, in principle applicable to all languages (i.e. a de facto standard) The common definition will be based on the experience with different languages, and will prevent reinvention of wheels The common definition will ensure interoperability and interconnectivity (especially for multilingual or cross-lingual applications)
Riga, Apr Other BLARK benefits Experience from other languages will help making cost estimations Adoption of a BLARK common to all languages may help in persuading funders to support the creation of the BLARK Adoption of a common BLARK may facilitate porting between languages
Riga, Apr To conclude Further development of Baltic HLT is a necessity for your society and an opportunity for your ICT Defragmentation of R&D efforts is a must Strengthen your position in Europe Ensure optimal conditions for the development of human and language resources for your region
Riga, Apr Thank you!