NewswireToday - /newswire/ -
Amsterdam, North-Holland, Netherlands, 2012/09/07 - TAUS, the translation innovation think tank and platform for industry-shared services, announces the launch of the Developing Talent Initiative, a collaborative project with universities around the world - TranslationAutomation.com.
This project allows students from all over the world to build MT engines using domain specific data from TAUS, supervised by the TAUS Labs team and an expert Advisory Committee. In the process, students gain a greater understanding of commercial requirements while industry also benefits from the experimentation undertaken by some of the brightest young computational linguists globally. The results are made publicly available on the TAUS site, helping to share learnings and inspire others.
The TAUS Data repository currently hosts over 50 billion words of trusted human translations in 2,200 plus language pairs. This represents the largest open source of quality language data to build and improve machine translation systems. The lessons that are learnt from experimentation using such a large volume of domain specific data have the potential to benefit business and society tremendously.
“Machine translation technology has lost its negative connotations”, says Jaap van der Meer, director of TAUS. “We see a rapid adoption of MT among buyers and providers of translation services. But most adopters are still at a loss at how they can improve the quality and train customized MT engines. The TAUS Developing Talent Project serves a double purpose. The industry will learn from the experiments undertaken by students in universities around the world. And academics will learn what the real world needs and they get access to the largest pool of human quality translation data.”
At launch fourteen universities in Asia, Europe and the Americas are taking part in the TAUS Developing Talent Initiative. Post-graduate students from these universities will take part in the process of data preparation, engines training/optimization/testing and translation quality evaluation. The TAUS Dynamic Quality Framework, the industry benchmark, will be used for quality evaluation and benchmarking of the MT engines.
“It is exciting to expand our virtual team of testers and MT specialists around the world so quickly,” says Maxim Khalilov, research director at TAUS and leader of the Developing Talent project. “Having worked in academia for 8 years, I know very well how crucial real translation data are to the success of MT research. Understanding the real world needs also helps universities to focus on research that makes the biggest difference to business and society. We look forward to working with many more universities and students around the world on this multi-year project.”
TAUS (tauslabs.com) is an innovation think tank and platform for shared services for the translation industry. TAUS supports entrepreneurs and principals in the translation industry to share and define new strategies through a comprehensive range of events, publications and knowledge tools.
Or contact Maxim Khalilov (maxim[.]tauslabs.com).