Advances in MT: What Is Exciting and Shows Promise Ahead?
Dion Wiggins & Philipp Koehn
Thursday, 19 September, 2019
This webinar provides an update to the webinar of the same name that Dion and Philipp presented in September 2018. In the last year there have been a number of advances in deep learning which are becoming visible to academia and available commercially. New approaches to machine translation that use neural architectures in novel ways are providing some very interesting and effective results. New neural architectures are proposed and ideas coming from such diverse fields as computer vision, game playing, and speech recognition can be applied to machine translation as well.
At the practical end, we are still learning about the deployment challenges of this technology, since old methods, for example, to integrate terminology databases or domain adaptation, no longer apply. There are new approaches to domain adaptation, data synthesis and data manufacturing. It is now common place to use tens or even hundreds of millions of bilingual sentences where just a few years ago a few million bilingual sentences was considered a large amount of data when building state-of-the-art MT engines.
This presentation gives an overview of the latest developments in research and what this means for practical deployment.
Dion Wiggins, Omniscien Technologies' CTO and Co-Founder, is a highly experienced ICT industry visionary, entrepreneur, analyst and consultant. He has comprehensive knowledge in the fields of software development, architecture and management, as well as an in-depth understanding of Asian ICT markets. He is an accomplished speaker and has a high media profile for his perceptive analysis of ICT in Asia Pacific. Previously Dion was Vice President and Research Director for Gartner based in Hong Kong, where his research reports on ICT in China had a crucial impact on how the world views this market. Dion is also a well-known pioneer of the Asian Internet Industry, being the founder of one of Asia's first ever ISPs (Asia Online in Hong Kong). In his role as consultant, Dion advised literally hundreds of enterprises on their ICT strategy.
Philipp Koehn is Chief Scientist at Omniscien Technologies and Professor of Computer Science at Johns Hopkins University; he also holds the Chair for Machine Translation in the School of Informatics at the University of Edinburgh. Koehn is a leader in the field of statistical MT research with over 100 publications. He is the author of the textbook in the field. Under his leadership the open source Moses system has become the de-facto standard toolkit for MT in research and commercial deployment. Koehn led international research projects such as Euromatrix and CASMACAT; his research has been funded by the European Union, DARPA, Google, Facebook, Amazon, Bloomberg, and several other funding agencies. Koehn received his PhD in 2003 from the University of Southern California and was a postdoctoral research associate at MIT. He was a finalist for the European Patent Office’s European Inventor Award in 2013 and received the Award of Honor from the International Association of Machine Translation in 2015. At Omniscien Koehn refined MT technology for use in real-world deployments and helped develop methods for data acquisition and refinement. Koehn continues to drive innovation and technological development at Omniscien.
Most machine translation output is fine for basic Western LSP needs, as they usually translate within European...
Machine translation is great when it works, but it fails often and in unpredictable ways. What causes bad machine...