The article has been co-written by Elia Yuste, Alexandre Helle and Manuel Herranz from Pangeanic and by Hirokazu Suzuki from the Corporate Research & Development Center of Toshiba Corporation.

It describes the efforts and progress made with the statistical machine translation open-source platform Moses and Toshiba’s rule-based system to obtain better outputs.

A copy of the article is available here Pangeanic_Toshiba_Hybridation_AAMT

Pangeanic’s development team gathered important input on advances on machine translation hybridation from leading academic at the International Workshop on Using Linguistic Information for Hybrid Machine Translation (LIHMT-2011) and the practical Saturday session ML4HMT (META NET WP2) in conjunction with DKFI. The sessions were geared for development and R&D personnel and those with a personal interest in making use of the best of research on MT and its different flavours to improve current state-of-the-art systems. Attendants and presenters were leading academics from the US, European Union and Japan involved in different MT areas. The theme was state-of-the-art developments in combinations (often involving Moses) and hybridation of rule-based approaches with statistics. Sessions dealt with combined approaches using syntax, grammatical information, rules and statistical systems.

As different research teams are facing the same problems worldwide, some similar, some new some quite imaginative approaches are beginning to emerge, for example:

• Lemmatisation, annotation for morphologically-rich languages, for example Czech and Basque and even lesser resources in the case of the 2nd one.
• Syntax-based approaches and word re-ordering for very unrelated languages (such as Asian or Semitic languages into and out of European languages)
• Web-based annotation tools
• Hybridisation of techniques, starting from analysis at a morphological layer, then analytical layers, tectogrammatical layers, and then transfer, and on to synthesis to t-layers, a-layer and m-layer.
• Word disambiguation
• Mixture of rule-based and statistical approaches to improve predictability.
• Post-editing effort estimation for MT systems and with systems including no linguistic features or having some. Linguistic features are relevant for direct useful error detections and for automatic post-editing. But for sentence-level CE there are issues with sparsity and with representation (length bias).
• New metrics like VERTa, using linguistic knowledge organised in different levels (lexical, morphological, syntactic information and sentence semantics)

Carnegie Mellon University's Alon Lavie and Manuel Herranz exchanging views on hybridation

Manuel will be speaking about the application of machine translation technologies in EU languages for localization and also the new development for EN –> JP. This is geared for LSPs and corporations which need to deal with sensitive data and create, build and maintain their own machine translation technologies in-house. With more and more security breaches, having the ability to offer not only TM but also machine translation pre-translation in-house and a post-editing environment is incresingly becoming a necessity for many companies which deal with sensitive data.

Machine translation has been applied very successfully for many EU languages at Pangeanic and within specific fields. Pangeanic builds engines for particular domains like automotive, bio & medicine and very interestingly for Japanese industry, consumer and professional electronics. Any field which has a “controlled input”  is ideal for automation.  Moreover, our latest offering – the famous “DIY SMT”- provides the ability to re-train engines and create new domains in-house, with no data transfer. This gives full ownership and control to LSPs and corporations alike.

This year’s presentation follows Pangeanic’s first introduction of SMT techniques for LSPs in 2010 (pictures below)

Manuel at BIJ-Pangeanic's booth

Iwanaga-san busy explaining the benefits of Statistical MT for LSPs

Pangeanic / PangeaMT were pleased to cooperate with the Localization World Organizing Committee and became Gala Dinner Sponsors, an event attended by over 400 international delegates of a total 550 attending this conference edition, the largest in number of participants and one of the most successful editions to date.

The PangeaMT booth was visited on a constant basis by well-known localization industry players as well as newcomers to the industry, established fellow LSP companies in Europe, the US, South America, Africa and Asia-Pacific, and last but not least, freelance translation specialists and consultants operating internationally.

This was an extraordinary setting to discuss collaboration and representation avenues, sales opportunities and most importantly, demo our PangeaMT DIY technology while first analyzing customized translation automation needs with our interlocutors.

Apart from exhibiting PangeaMT, also with a focus on the post-editing and  other top-notch linguistic services that  Pangeanic may offer, Manuel  Herranz and Elia Yuste also took part in the  Conference program as  speakers at the E7 track hosted by Jaap van der  Meer (TAUS) and entitled  MT Experiences at IBM, Sony Europe and  Sybase. Fausto Prastaro (Sony UK) and Elia Yuste and Kerstin Bier (Sybase, a SAP company) and Manuel  Herranz presented their two  respective use cases. Presentations available upon request.

We are looking forward to seeing Localization World happening in Spain again in future spring time editions and are happy to confirm our participation to the next edition of the Conference in Santa Clara, Silicon Valley, California, in October’11.

]]> http://www.pangea.com.mt/en/2011/07/proactive-and-highly-successful-participation-of-pangeanicpangeamt-in-localization-world-barcelona-2011/feed/ 0 http://www.pangea.com.mt/en/2011/05/pangeanic%e2%80%99s-ceo-to-speak-at-the-elia%e2%80%99s-nd-in-stockholm-tomorrow/ http://www.pangea.com.mt/en/2011/05/pangeanic%e2%80%99s-ceo-to-speak-at-the-elia%e2%80%99s-nd-in-stockholm-tomorrow/#comments Fri, 13 May 2011 14:46:33 +0000 admin http://www.pangea.com.mt/en/?p=834 Continue reading →]]> Manuel Herranz, Pangeanic’s CEO will be the invited speaker closing the international conference entitled ELIA’s Networking Days Stockholm on 14th May.

ELIA stands for the European Language Industry Association. Their conference, now in its 10^th anniversary, is a respected forum of our industry that has gathered momentum in the last editions, particularly last year’s in Dublin. We at Pangeanic are excited about speaking at ELIA this year, where we are happy to meet up with familiar faces, get to know other fellow industry players and, above all, enthusiastically encourage the adoption of democratizing, customizable and user-centric MT technologies, such as PangeaMT.

Manuel’s talk is entitled Moses Customization for LSPs – the Pangeanic experience.  Click here to read the abstract.

You can learn more from ELIA here.

The prestigious Spain Technology web portal features an article on PangeaMT-BIO.

It highlights this as a case where language technologies, in particular, the customized, in-domain Machine Translation solution range from Pangeanic, are at the service of the Biotechnology sector.

If you are a company or institution belonging to this industry vertical and have a steady or growing need for translating a significant amount of content, please do let us know. We will offer you a non-obligation consulting conversation aimed at analysing your multilingual communication needs and devising the best translation automation strategy in connection with your in-domain BIO scientific literature and collateral materials.

An English-Spanish PangeaMT-BIO system (non-customized) will be exhibited at the International BIO Convention in the week of 27th June 2011 in Washington, D.C., USA.

Valencia, Spain, April 27, 2011 — Elia Yuste, Pangeanic’s BDM to present PangeaMT’s latest machine translation offering in the BIO 2011 Conference in Washington DC at the end of June.

PangeaMT, Pangeanic’s language technology division, is a European leader in the tailored implementation of independent, open source statistical translation automation solutions for specialized domains. A growing number of global companies in need of establishing and consolidating their international presence have trusted us their multilingual content production strategy. Translation automation in scientific and technical domains is not only technologically possible but commercially necessary. However, not any freely available, open-domain translation engine should be used. Think also about your corporate confidentiality policies. It is thus advisable to deploy translation solutions that are data-driven, domain-specific and fully customized in accordance to the corporate literature and knowledge of the client commissioning the PangeaMT engine.

Corporations, content writers and other language service providers (LSPs) have benefited from this approach. Even if you translate just half a million words a year, you are a prime candidate to benefit from this technology. PangeaMT brings together a team of professionals with expertise in computational linguistics, translation, content and knowledge management, with the flexibility to adapt to your needs and guide you through each step. As part of our consultancy and customized engine creation, the PangeaMT team can look after the hard work involved in data mining, collection and analysis. If you already have bilingual text files, we will leverage and enhance your existing translation databases and if need be, add content that is representative of your domain.

PangeaMT’s mission is to bring democracy to the machine translation world and make the technology affordable and usable for as many users as possible belonging to highly specialized fields. We have been serving leading industry vertical players from the automotive, software, engineering or consumer electronic fields, to name a few. “It is about time that we started to make significant inroads in the biotechnology arena as well, a rapidly growing sector that will represent Spain’s 3% GDP by 2030. Thanks to the PangeaMT-BIO system for the English-Spanish language pair, we will be able to help Spanish-speaking biotech organizations to firmly step in the international biotechnology scene, thanks to properly translated content that derives from the latest statistical and hybrid translation technologies with the human linguistic touch,” Manuel Herranz, Pangeanic’s CEO says.

Translation industry consultant and expert Don DePalma from Common Sense Advisory agrees with Elia Yuste, Pangeanic’s Business Development Manager that the BIO International Convention in Washington DC in late June is the perfect scenario to exhibit PangeaMT-BIO in the US. Biotechnology players that wish to internationalize their products and services by sharing information about them in languages other than English will surely find the concept behind PangeaMT-BIO and Pangeanic’s commitment to customized and cost-effective translation solutions most attractive.

Pangeanic, the 15-year old, Spanish leading provider of translation, publication, software and web localization and manual-making services, works independently and in association with sister offices in Japan and China for (cross-national) institutions, famous multinationals such as Sony or Sybase-SAP, government agencies and other language companies all over the world. Our built-in quality service philosophy is accompanied by the latest computer applications and state-of-the-art language quality assurance (QA) technology.

Our whole package is engineered to speed and reduce your localization and translation costs in any language. Our work ethos is directed to make the language translation process as smooth, painless and cost-effective as possible. We believe in the automation of as many processes as possible. You reap the benefits of PangeaMT domain- and customer-specific machine translation technology in your internationalization and globalization effort.

If you are a Biotechnology player in need of specialized content translation you will love the new English<>Spanish PangeaMT-BIO engine.

# # #

Contact Person: Ms Elia Yuste, Pangeanic’s Business Development Manager and PangeaMT’s Scientific and Commercial Lead
Email:  e.yuste@pangeanic.com / eyuste@pangea.com.mt
Division’s Web Site & Blog: www.pangea.com.mt / pangeanic.wordpress.com/

Now, the ultimate release which will help you gain technological independence and build your own MT eco-system, updating it and making grow as you work. Use powerful, tested Moses in a new customized PangeaMT package to retrain and update your engines at will as if you were updating translation memories and create new domains.

Shorter is better, so here’s a brief list of the new features that will let you be in control of your own solution and make it grow in the direction(s) you wish:

* Build a training structure (domain or theme directories within a language: financial, technical, marketing, etc).

* Populate each directory with the bitext TMX files.

* Choose the incremental retraining procedure (time or more data). Your engines will keep on using your new data and improving translation results.

* Create new combination engines at will “engineering + financial” for mixed domains.

* Feed your own data or add more data from other sources: you are your own master to personalize each engine for each of your clients or for a vertical domain.

* Receive translation data as TMX, XLIFF, txt, ttx, etc.

* Post-edit using the tool of your choice. Leverage TM and MT.

Components & benefits

* Stats Control Panel – all the data you have in your system: number of words, sentences, scores.

 ¹As reported in AMTA's last conference and acknowledged by EuroMatrix Plus.

We append the conference programme below. Please contact us if you need further details or would like to receive a copy of Manuel’s presentation later.

EUROPEAN COMMISSION

Directorate-General for Translation

Directorate R – Resources,

R.3 – Informatics

TRANSLATORS AND MACHINE TRANSLATION

DGT CONFERENCE, 6 APRIL 2011, G-12 3/21– JMO M6

Programme

14.15 Welcome and introduction

by Mr Piet Verleysen, Director for Resources, DGT

14.30 Machine translation technology – introduction

by Dr Andreas Eisele, Machine Translation Project Manager, Informatics unit, DGT

15.10 Machine Translation and translators: evidence from research

by Dr Sharon O’Brien, Lecturer in Translation Studies, School of Applied Language and Intercultural Studies, Dublin City University, Ireland

15.50 Machine translation: the perspective of the language industry

by Mr Manuel Herranz, owner & senior strategy officer of the “Pangeanic” translation company, Spain

16.30 Using machine translation in DGT: the experience and the perspective of a language department

by Mr Hilário Leal Fontes, translator, Portuguese Language Department, DGT

17.00 Discussion

17.30 Closing

by Mr Josep Bonet, Head of the Informatics Unit, DGT

This Workshop is linked to the EU-funded Multilingual Web project geared at establishing best practices and standards related to all aspects of creating, localizing and deploying the Web multilingually.

Manuel’s talk is entitled Open Standards in Machine Translation. The abstract is as follows:

The web is an open space and the standards by which it is “governed” must be open. However, one barrier clearly remains to make the web even more transnational and truly global. This has been called “the language barrier”. Language Service Providers translation business model is clearly antiquated and it is increasingly being questioned when we face real translation needs by web users. Here, immediacy is paramount. This talk is about open standards in machine translation technologies and workflows, supporting a truly multilingual web.

More information about the rest of speakers’ contributions and the event can be found here.

If you could not make it to Pisa, you can still follow the Workshop bymeans of the video streaming option or through Twitter.