Fostering Language Resources Network K h FLa Re

Fostering Language Resources Network K h FLa. Re a Net l i d C h o u k r i E L R A The European Language Resources and Technologies Forum: Shaping the Future of the Multilingual Digital Europe Session 1 Broadening the Coverage, Addressing the Gaps E/BLARK as a tool for Language Resources Coverage assessment, Road mapping, and Language Policy planning Some thoughts and considerations Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net

Fostering Language Resources Network ELRA's Views on WHAT ARE THE MISSING LRs PIECES FLa. Re Net Ø Ø Derived from Cataloguing q What is "NOT" available at "fair conditions" § (Pricing, licensing, re-usability, etc. ) q What exist but is not available (traded vs not traded) , see next slide q What exist but is not (yet? ) identified (assumptions) q What is lost !! Not archived (even what have been funded by public agencies) q What does not exist (e. g. is really missing, our guesses) Related to a number of Dimensions to consider: • • Technologies/Dev : ASR, TTS, S 2 S, MT Languages (& Varieties) Domains Technologies/Evaluation n ELRA ICom (Infrastructure committee) & Universal Catalogue & related issues (Metadata, New Data licensing, distribution/sharing mechanisms, etc. ) Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 2

Fostering Language Resources Network FLa. Re Net Some input from Bain market analysis (for ELRA) Not available Spending on LRs is driven by: Not traded • New applications • Number of languages • Public Policies • Decreasing number of new customers Traded Share of publicly traded LRs will be driven by: • Availability • Differentiation rationales Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 3

Fostering Language Resources Network BLARK as an Instrument for Policy Making and Policy Makers FLa. Re Net n What can you do for BLARK? q q n What can BLARK do for you? q n Define, Specify, Improve, etc. see the ELRA site: www. blark. org Help enhance the content of the Universal Catalogue Help you with useful input for your R&D and your Cooperations, What can your "agency" do for BLARK q q q Ensure that it is accurate and reflects community plans/trends OR some "body" legitimate strategy. Use it as a "consistent" planning/roadmapping instrument Use it as a "State of the Art" Inventory to understand the "country" position within HLT sector Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 4

Fostering Language Resources Network WHAT CAN WE (WE? ) DO TO ADDRESS SUCH SHORTAGES in LRs FLa. Re Net n Universal Catalogue and Common metadata (e. g. ELRA/LDC/NICT/OLAC): q n How can this be a LRs "You. Tube" or the like? Revise Distribution mechanisms & pricing policies q e-Licensing & e-distribution, and other frameworks to foster the successful sharing of LRs? q freeware, shareware, "fair" prices Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 5

Fostering Language Resources Network FLa. Re Net n ENHANCE SYNERGIES IN LR PRODUCTION (OR PACKAGING) THROUGH COOPERATION BETWEEN COMMUNITIES Check who is producing, as "side-products", essential LRs for HLT n e. g. Publishers, broadcast companies, etc. Public data (Parliaments (Euro. Parl), governments, agencies, etc. ) n "Research Fair Act" a la European Union n ? ? Flare. Net – Forum Vienna 12 -13/02/2009 ELRA to launch a Petition on this…. . Simplify the current legal framework Session 1/KC/ELRA 1 FLa. Re. Net 6

Fostering Language Resources Network FLa. Re Net CAN WE LEAVE SOME OF THIS TO "LUCRATIVE BUSINESS PRINCIPLES" n Sure, many marketplaces are left to private initiatives e. g. Banking & trading n See S 6 For Market aspects… n What happen when Private sector produces LRs (even with Public funds) ? q n (20%) Buy, (25%) make and share, (55%) make & keep (Bain) What would be the costs to regain existing positions if things develop badly? u Need for a European LRs Investment FUND (Contributors, ROI) Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 7

Fostering Language Resources Network FLa. Re Net HOW CAN LR DEVELOPED TODAY …. LAST FOR EVER!! n Usability issue & Availability (inc. Accurate Metadata) n Standards/best practices / Validation of the quality n Interoperability n Sustainability /Bug reporting/ Updates n Above all: excellent specifications, consideration of today's state of the art and today's trends n e. g. Broadcast news …. . Audio tracks … Video, Audio: multiple tracks, multi-track Subtitles, Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 8

Fostering Language Resources Network FLa. Re Net n HOW CAN THESE CONSIDERATIONS BE LINKED TO BLARK/ELARK: BLARK & ELARK are only instruments (what to do and not "the do") q q ELRA efforts… ICom (Infrastructure Com. catalog, e-distrib, e-license, e-share) q n Universal Catalogue (Identified all existing resources, worldwide) ……. They are instruments that helps conduct: n Coverage assessment & Identification of Gaps, n Road mapping & HLT Policy planning n But also R&D & Cooperations between researchers (BLARK reflects "state of the art" ) n (enhanced) coordination between all involved agencies …. If n used adequately They require accurate design/completion, Regular updating and extension Flare. Net – Forum Vienna 12 -13/02/2009 Session 1/KC/ELRA 1 FLa. Re. Net 11