Keynote Speakers

[ top ]

At Symantec, we have been using MT with TM in localization for several years. Our approach has focused on optimizing our customized SYSTRAN MT system and on the post-editing of MT output by external vendors. With MT, we have increased the speed and consistency of translation, which also enabled us to process larger document sets within shorter turnaround times.

One of the main challenges in having MT output post-edited is being able to meet ever increasing quality expectations, especially from a terminology perspective. We refine our MT output with several layers of customization (such as User Dictionaries, Translation Stylesheets and Post-processing) that will be discussed in this presentation. Despite substantial quality gains with these techniques, we continue the search for ways to minimize post-editing effort. We present new technology based on SYSTRAN's hybrid approach, and focus on the technical and linguistic challenges of integrating it into our existing MT workflow.

Johann Roturier has been working in the Localization industry for the last six years, focusing on the evaluation, deployment, and integration of new language and translation technologies (such as MT) into localisation workflows. He obtained his Ph.D from Dublin City University in 2007 in the field of controlled language and machine translation usability. His current research interests include controlled authoring, terminology harvesting and automated post-editing. He is also a member of the XLIFF Technical Committee and of the Centre for Next Generation Localisation (CNGL).

[ top ]

Human Translation is a $12 billion a year market of which professional translators get a significant share. Helping translators to be 10% more productive can create a potential market bigger than the current MT market. Is this a realistic direction? What are the requirements of the human translation market? This presentation will provide a live demo of MyMemory, the world's largest translation memory archive - a practical approach to combining machine translation and human translation.

Marco Trombetti founded Translated in 1999 and serves as its CEO. Translated is a web based localization company that offers language services to over 9.000 customers. Marco started his career developing IT solutions for the digital mapping industry and then worked as a consultant in the field of artificial intelligence, specializing in information retrieval and natural language processing. Marco is a serial entrepreneur and serves as a director at various technology companies. He studied physics at the University of Rome.

[ top ]

This talk will outline recent research developments in statistical machine translation in a way that will be comprehensible not only to researchers in the field, but to interested developers and users. There will be a particular focus on themes covered by papers presented at the conference. Some of the salient themes are:

Dr Pierre Isabelle is Principal Scientist and manager of the Interactive Language Technology Group at the National Research Council. He has devoted his career to doing and managing research in natural language technology at the University of Montreal (TAUM, RALI), in public research labs (CITI, NRC) and with Xerox (Grenoble). He is internationally known for his accomplishments in that research area, especially in machine (-aided) translation.

Dr. Roland Kuhn is a Research Officer in the Interactive Language Technology Group of the National Research Council of Canada (NRC). He has carried out research in both automatic speech recognition (for the Centre de recherche informatique de Montréal and for Panasonic Speech Technology Laboratory / Santa Barbara), and in machine translation (for NRC). He has authored 44 refereed publications and holds 29 US patents.

[ top ]

The Director of the Foreign Language Program Office, Dan Scott, will present his viewpoints on users, machine translation needs, and measurement of success. He will present a range of user scenarios which present challenges to the MT community to create valid and replicable measures of success. These metrics must be meaningful within the context of various workflows and use cases. If there is no single "MT solution" nor a single metric, then the riddle for the MT community to answer is: For users, what exactly is the meaning of 94% success?

[ top ]

This talk will discuss machine translation evaluation in the context of multilingual speech and text processing applications, including the following highlights: (1) creation of new machine translation evaluation methods linked with standard Defense Language Institute measures of human foreign language skills, to measure the utility of machine translation for government applications; (2) development of a testbed for machine translation algorithms, with a special focus on less commonly taught languages. The key idea is to use human translation practices as the control condition for a baseline, and to contrast test scores on that condition with conditions involving human interaction with machine translation systems. Higher standardized test scores means better translation.

Douglas Jones is a technical staff member at MIT Lincoln Laboratory in the Information Systems Technology Group. He joined Lincoln in 2001 and worked to establish a new research program on machine translation with an emphasis on adapting standardized tests for technology evaluation. His civil service job at the Department of Defense Natural Language Research Branch set the stage for his current focus. He earned bachelor.s and master's degrees in linguistics from Stanford University and a doctoral degree in linguistics from MIT, supervised by Noam Chomsky.