<div dir="ltr"><div class="gmail-blog__content" style="box-sizing:border-box;outline:0px;padding:0px;border-bottom:0px;overflow:hidden;width:599.391px;color:rgb(0,0,0);font-family:Lato,sans-serif;font-size:16px"><div class="gmail-blog__title" style="box-sizing:border-box;outline:0px;margin:0px;padding:0px"><h4 style="box-sizing:border-box;outline:0px;margin:0px 0px 15px;color:rgb(53,53,54);line-height:1.2;font-size:2rem;padding:0px;letter-spacing:-1px">EU Awards USD 4.8m in Neural MT Grants for Low-Resource Languages</h4></div><div class="gmail-blog__meta gmail-d-none gmail-d-md-block" style="box-sizing:border-box;outline:0px;line-height:1;margin-top:0px;padding:10px 0px"><div class="gmail-date_ago" style="box-sizing:border-box;outline:0px;font-size:12px;letter-spacing:0.5px;display:inline-block;line-height:20px;color:rgb(153,153,153);text-transform:capitalize"><span class="gmail-type gmail-pcolorbg elapsed-time" style="box-sizing:border-box;outline:0px;background:rgb(45,193,230);border-radius:15px;color:rgb(255,255,255);font-weight:700;line-height:20px;margin-right:10px;padding:1px 15px">6 Hours Ago</span></div> <div class="gmail-author" style="box-sizing:border-box;outline:0px;font-size:12px;letter-spacing:0.5px;display:inline-block;line-height:20px;text-transform:capitalize"><a href="https://slator.com/uncategorized/" class="gmail-blog__category" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);display:inline-block;font-weight:700">Uncategorized</a> <span class="gmail-dot" style="box-sizing:border-box;outline:0px;color:rgb(153,153,153);display:inline-block;line-height:20px;margin:0px 5px">·</span> <p class="entry-author" style="box-sizing:border-box;outline:0px;margin-bottom:0px;margin-top:0px;color:rgb(137,143,150);line-height:20px;margin-right:0px;display:inline-block;font-weight:700">By <a href="https://slator.com/author/ginodino/" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230)">Gino Diño</a></p> <div class="gmail-date_time" style="box-sizing:border-box;outline:0px;display:inline-block;line-height:20px"><p style="box-sizing:border-box;outline:0px;margin-bottom:0px;margin-top:0px;color:rgb(137,143,150);line-height:20px;margin-right:0px;display:inline-block;font-weight:700">On March 25, 2019</p></div></div></div></div><div class="gmail-single_blog_content" style="box-sizing:border-box;outline:0px;padding:0px;overflow:hidden;width:599.391px;color:rgb(0,0,0);font-family:Lato,sans-serif;font-size:16px"><img class="gmail-inline-post-thumbnail gmail-d-none gmail-d-md-block" src="https://slator.com/assets/2019/03/nmteugrant4march25.jpg" alt="EU Awards USD 4.8m in Neural MT Grants for Low-Resource Languages" style="box-sizing: border-box; outline: 0px; border-style: none; vertical-align: middle; max-width: 360px; height: auto; margin-bottom: 0px; border-radius: 0.75rem; float: left; margin-right: 1.45rem; margin-top: 0.5rem; width: 281.703px; display: block;"><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">On February 27, 2019, the Connecting Europe Facility Telecommunications (CEF Telecom) sector released the list of <a href="https://ec.europa.eu/inea/en/news-events/newsroom/cef-telecom-almost-%E2%82%AC10.5-million-eu-investment-to-boost-european-digital" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">selected proposals</a> in the areas of Automated Translation, eDelivery, and eInvoicing for their 2018 work program. A total of over EUR 4.3m (USD 4.8m) has been earmarked for five separate project proposals on Automated Translation.<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">CEF Telecom opened a <a href="https://slator.com/demand-drivers/eu-to-grant-eur-5m-for-low-resource-language-data-and-mt-implementation/" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">call for proposals in July 2018</a>, making available a maximum grant of EUR 5m per selected project. The call for proposals aimed to further improve the capability of the European Union’s neural machine translation (NMT) facility, <a rel="noreferrer noopener" href="https://ec.europa.eu/cefdigital/wiki/display/CEFDIGITAL/What+is+eTranslation+-+Overview" target="_blank" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">eTranslation</a>, in translating low-resource languages.<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">Selected proposals are expected to implement NMT engines, collect more training data for high-priority, low-resource languages, and assist in the integration of the eTranslation facility into the EU’s other online services that may require multilingual functionality.</p><div class="gmail-single_blog_ad_top" style="box-sizing:border-box;outline:0px;clear:both;margin-bottom:28px;overflow:hidden;max-width:100%"><div class="gmail-ad-unit" style="box-sizing:border-box;outline:0px;height:0px;padding-bottom:69.9219px;width:599.391px;max-width:100%"><span class="gmail-ad-unit-title" style="box-sizing:border-box;outline:0px;color:rgb(230,230,230);display:block;font-size:10px;height:20px;line-height:20px;text-align:center;text-transform:uppercase;width:90px;max-width:100%">ADVERTISEMENT</span><div class="gmail-ad-unit-container" style="box-sizing:border-box;outline:0px;height:69.9219px;margin:0px;width:599.391px;background:rgb(124,199,154);max-width:728px"><div id="gmail-placement3322850" style="box-sizing:border-box;outline:0px;height:69.9219px;margin:0px;width:599.391px;max-width:100%"><div id="gmail-served_332285_0" style="box-sizing:border-box;outline:0px;height:69.9219px;margin:0px;width:599.391px;max-width:100%;padding:0px"><a href="https://adbutler-fermion.com/redirect.spark?MID=167169&plid=886293&setID=332285&channelID=0&CID=0&banID=519664292&PID=0&textadID=0&tc=1&mt=1553526394316109&sw=1366&sh=768&spr=1&hc=3a7efe891fe31b8df88ce43f145c5eadfc86b8aa&location=" target="_blank" rel="nofollow" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%"><img src="https://adbutler-fermion.com/getad.img/;libID=610783" alt="" title="" border="0" width="728" height="90" style="box-sizing: border-box; outline: 0px; border-style: none; vertical-align: middle; max-width: 100%; height: 69.9219px; margin: 0px; width: 599.391px;"></a></div></div></div></div></div><h3 style="box-sizing:border-box;outline:0px;margin:0px 0px 1.2rem;line-height:1.2;font-size:24px;padding:0px;max-width:100%">Automated Translation</h3><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">CEF Telecom reportedly received 35 project proposals in response to the call and ultimately selected five for Automated Translation, three for eDelivery, and nine for eInvoicing. The total proposed maximum funding allocated for all 17 exceeds EUR 10.3m (USD 11.6m).<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">The five project proposals selected for the Automated Translation section of the work program and the maximum amount of funding earmarked for each are as follows:<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><table class="gmail-wp-block-table gmail-has-fixed-layout gmail-is-style-regular" border="1" style="box-sizing:border-box;outline:0px;border-collapse:collapse;min-width:100%;margin-bottom:15px;width:599px;table-layout:fixed;max-width:100%"><tbody style="box-sizing:border-box;outline:0px;max-width:100%"><tr style="box-sizing:border-box;outline:0px;max-width:100%"><td style="box-sizing:border-box;outline:0px;max-width:100%">Proposal</td><td style="box-sizing:border-box;outline:0px;max-width:100%">Coordinated by</td><td style="box-sizing:border-box;outline:0px;max-width:100%">Recommended Funding (EUR)</td></tr><tr style="box-sizing:border-box;outline:0px;max-width:100%"><td style="box-sizing:border-box;outline:0px;max-width:100%">PRINCIPLE – Providing Resources in Irish, Norwegian, Croatian and Icelandic for Purposes of Language Engineering</td><td style="box-sizing:border-box;outline:0px;max-width:100%">Dublin City University</td><td style="box-sizing:border-box;outline:0px;max-width:100%">1,138,781</td></tr><tr style="box-sizing:border-box;outline:0px;max-width:100%"><td style="box-sizing:border-box;outline:0px;max-width:100%">OCCAM</td><td style="box-sizing:border-box;outline:0px;max-width:100%">Brno University of Technology</td><td style="box-sizing:border-box;outline:0px;max-width:100%">973,894</td></tr><tr style="box-sizing:border-box;outline:0px;max-width:100%"><td style="box-sizing:border-box;outline:0px;max-width:100%">EuroPat: Unleashing European Patent Translations</td><td style="box-sizing:border-box;outline:0px;max-width:100%">University of Edinburgh</td><td style="box-sizing:border-box;outline:0px;max-width:100%">695,890</td></tr><tr style="box-sizing:border-box;outline:0px;max-width:100%"><td style="box-sizing:border-box;outline:0px;max-width:100%">Continued Web-Scale Provision of Parallel Corpora for European Languages</td><td style="box-sizing:border-box;outline:0px;max-width:100%">University of Edinburgh</td><td style="box-sizing:border-box;outline:0px;max-width:100%">889,649</td></tr><tr style="box-sizing:border-box;outline:0px;max-width:100%"><td style="box-sizing:border-box;outline:0px;max-width:100%">Translation Automation Services for EU Council Presidency</td><td style="box-sizing:border-box;outline:0px;max-width:100%">Tilde</td><td style="box-sizing:border-box;outline:0px;max-width:100%">620,800</td></tr></tbody></table><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">Four language service and technology providers are involved in the listed projects: Iconic Translation Machines, Prompsit, Omniscien Technologies, and Tilde.<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">PRINCIPLE is a consortium among Dublin City University’s ADAPT Centre, Dublin-based MT provider Iconic Translation Machines, the National Library of Norway, the University of Zagreb, and the University of Iceland. According to an <a href="https://iconictranslation.com/2019/03/iconics-consortium-awarded-contract-by-eu-for-over-e1-million-euro-for-neural-mt/" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">article on the Iconic website</a>, the consortium partners “will collaborate closely with government agencies, national public administration bodies and data holders in Croatia, Iceland, Ireland and Norway to ensure that the pooled language resources appropriately cover the technology needs.”<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">The “Continued Web-Scale Provision of Parallel Corpora for European Languages” proposal is a <a href="https://www.dlsi.ua.es/eines/noticia.cgi?id=eng&idn=420" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">continuation of previous, similarly titled, projects</a> by the same consortium, which includes the University of Edinburgh, the Universitat d’Alacant, language industry organization TAUS, Spain-based ICT company Prompsit Language Engineering, and Singapore-headquartered language technology company Omniscien Technologies.<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">Meanwhile, Latvian MT company Tilde is responsible for <a href="https://www.tilde.com/news/official-website-2018-eu-council-presidency-integrates-tildes-neural-mt-service" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">Translation Automation Services for the EU Council Presidency</a>, which provides the eTranslation platform with NMT for Bulgarian.</p><h3 style="box-sizing:border-box;outline:0px;margin:0px 0px 1.2rem;line-height:1.2;font-size:24px;padding:0px;max-width:100%">Millions of Euro into MT</h3><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">Launched in <a href="https://ec.europa.eu/info/resources-partners/machine-translation-public-administrations-etranslation_en" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">November 2017</a>, the eTranslation facility is the NMT replacement of the EU’s previous statistical MT solution called MT@EC. CEF said eTranslation is “trained using the vast Euramis translation memories, comprising <a rel="noreferrer noopener" href="https://ec.europa.eu/cefdigital/wiki/display/CEFDIGITAL/Machine+Translation" target="_blank" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">over 1 billion sentences in the 24 official EU languages</a> produced by the translators of the EU institutions over the past decades.” Euramis stands for “European advanced multilingual information system.”<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">MT plays a pivotal role in the EU’s efforts to overcome language barriers in its vision of a <a href="https://slator.com/demand-drivers/eu-spends-another-eur-2-4m-on-translation-research-for-digital-single-market/" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">Digital Single Market</a>. According to the <a rel="noreferrer noopener" href="https://ec.europa.eu/digital-single-market/en/language-technologies" target="_blank" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">Language Technologies policy</a> of the EU’s Digital Single Market strategy, current MT solutions available on the market “usually don’t reach the required levels of quality, or only for [a] limited number of languages, text types or topics.”<br style="box-sizing:border-box;outline:0px;max-width:100%"></p><p style="box-sizing:border-box;outline:0px;margin-bottom:28px;margin-top:0px;font-size:1rem;line-height:2rem;max-width:100%">To address this, Horizon 2020 — the biggest EU research and innovation program with a budget of EUR 80bn available from 2014 to 2020 — has been funding various MT undertakings, such as a <a href="https://slator.com/academia/the-european-union-takes-another-shot-at-machine-translation/" target="_blank" rel="noreferrer noopener" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">EUR 3m project</a> for the automated translation of Massive Open Online Courses.<br style="box-sizing:border-box;outline:0px;max-width:100%">At the same time, CEF “addresses the additional challenge of creating a complete infrastructure for language resources and processing tools.” CEF has funded efforts such as a <a rel="noreferrer noopener" href="https://slator.com/deal-wins/europe-awards-eur-5-8m-to-tackle-digital-markets-language-problem/" target="_blank" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">EUR 5.8m initial tender</a> to build a standalone Automated Translation service building upon MT@EC (at the time), a <a rel="noreferrer noopener" href="https://slator.com/deal-wins/eu-spends-eur-1-9m-to-customize-mt-for-state-and-regional-authorities/" target="_blank" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">EUR 1.9m project</a> to customize MT for state and regional authorities, a <a rel="noreferrer noopener" href="https://slator.com/demand-drivers/eu-spends-another-eur-2-4m-on-translation-research-for-digital-single-market/" target="_blank" style="box-sizing:border-box;outline:0px;background-color:transparent;text-decoration-line:none;color:rgb(45,193,230);max-width:100%">EUR 2.4m research project</a> for automated translation and, now, this EUR 4.3m for low-resource languages.</p></div><div><br></div>-- <br><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature">=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+<br><br> Harold F. Schiffman<br><br>Professor Emeritus of <br> Dravidian Linguistics and Culture <br>Dept. of South Asia Studies                     <br>University of Pennsylvania<br>Philadelphia, PA 19104-6305<br><br>Phone:  (215) 898-7475<br>Fax:  (215) 573-2138                                      <br><br>Email:  <a href="mailto:haroldfs@gmail.com" target="_blank">haroldfs@gmail.com</a><br><a href="http://ccat.sas.upenn.edu/~haroldfs/" target="_blank">http://ccat.sas.upenn.edu/~haroldfs/</a>    <br><br>-------------------------------------------------</div></div>