Sign up for our Webinars or listen to a Recording
“Practical Application of Neural Machine Translation”
Part of Omniscien Technologies webinar series (Free for all attendees) – Subscribe to our mailing list below to receive our invitation when registration is open
How to Webinar Series:
Language Studio 4
Technology Webinar Series:
Neural Machine Translation (NMT)
Introduction to Language Studio 4.0
The release of Language Studio™ V4.0 brings many new features which no other MT provider at present can match in terms of quality or productivity gains.
Designed for simplicity and ease of use while achieving the highest quality machine translation output, Language Studio™ V4.0 offers usage plans to meet any requirements or budget.
In this presentation, Dion Wiggins, Omniscien Technologies’ Chief Technology Officer, will walk you through the core features of the latest release of Language Studio™. Learn how to create custom engines and refine them over a short period to deliver best quality translation output.
This presentation covers:
- Translating with Language Studio
- Using Industry Engines
- Customizing Your Own Engines
- BASIC Customisation
- PROFESSIONAL Customisation
- Integrating with your TMS Platform and Editing Tools
View the recording of the webinar above.
NMT Webinar #2
Introduction to Neural Machine Translation (NMT)
This Omniscien Technologies webinar recording from January 24th, 2017 provides a comprehensive introduction to Neural Machine Translation (NMT) with comparisons to Rule Based Machine Translation (RBMT) and Statistical Machine Translation (SMT), it’s basic inner workings, as well as a view of both the opportunities that NMT offers as well as some of the early stage deployment challenges.
This webinar is the second webinar in the Omniscien Technologies webinar series on Neural Machine Translation (NMT).
Language Studio™ Engine Types
With the release of Language Studio V4.0 in Q3/16, Omniscien Technologies made available 3 new Machine Translation engine types. This webinar explains the differences between each engine type, what business scenarios and uses cases they are best suited for, and how to take advantage of each engine type. The three engines types supported and discussed on the call are:
· INDUSTRY MT Engines allow immediate access to hundreds of off-the-shelf engines that are ready to use without any customization. 13 industries ranging from Life Sciences, Automotive and Information Technology to eDiscovery are being deployed progressively across the 540 language pairs supported by Language Studio. Some control such as glossaries and non-translatable terms is available.
· BASIC Custom MT Engines are Do-It-Yourself (DIY) custom engines, built on top of an INDUSTRY engine as a foundation, providing a solid base to begin from for a custom engine. Users simply upload their training data (i.e. TMX, XLIFF, etc.) and submit their engines for automated training. This engine type allows for rapid customization when you have data and language assets suitable for training an engine with.
· PROFESSIONAL Custom MT Engines are an expert guided comprehensive customization that is based on a collaborative effort between Omniscien Technologies Language Studio Experts and the user, and expands of the features offered in a BASIC custom engine to deliver the highest possible translation quality. Each engine has a uniquely tailored Quality Improvement Plan that defines the steps that will be taken to customize the engine.
Put simply, this is similar to outsourcing the expert skills to the Language Studio team, while still keeping complete control of the project. There are many activities that are undertaken such as term management, normalization and development, in addition to data manufacturing and writing style control. With a Professional Custom MT Engine, it is even possible to develop a high quality custom engine when you have no data at all.
View the recording of the webinar above.
NMT Webinar #1:
The State of Neural Machine Translation
On November 24th, 2016, Omniscien Technologies organised a webinar with panel discussion on the State of Neural Machine Translation (NMT) with Prof. Philipp Koehn as the presenter.
The presentation and the panel discussion was focused on the current state of the technology development as well as the possibilities and challenges related to the deployment of NMT in relation to RBMT and SMT.
The slides are available for download here and the Frequently Asked Questions (FAQ) are available in the Omniscien Technologies Frequently Asked Questions web section. To playback the recoding see above.
The presenters and panelists were:
- Philipp Koehn holds a professorship in Computer Science at Johns Hopkins University and is the Chair for Machine Translation in the School of Informatics at the University of Edinburgh. Philipp Koehn is well-known as a leader in the field of statistical machine translation research, particularly in association with the open source Moses system, which has become the de-facto standard toolkit for machine translation in research and commercial deployment. Philipp is the author of numerous publications and he received the Award of Honor from the International Association of Machine Translation last year. As Omniscien Technologies’ Chief Scientist, Philipp refined machine translation technology for real-use and continuously drive innovation and technological development
- Kerstin Berns is founder of berns language consulting, based in Germany. She started her career as a translation system specialist for DaimlerChrysler, introducing machine translation to the DaimlerChrysler intranet in 1999 before working as team leader for the BMW Aftersales translation team. In 2005 Kerstin started her own company, an independent consultancy aiming at state-of-the-art quality and efficiency for the multilingual information process chain. Kerstin’s team caters to the needs of medium to large international companies with a strong interest in lean management and quality control. As an independent company, berns language consulting has no economic ties to any translation or software vendors whatsoever.
- Florian Faes is Co-Founder and Managing Director at Slator, which is an online publisher that makes business sense of the language services and technology market with news on the people and deals that shape the industry. Before launching Slator in 2015, Florian was with language service provider CLS Communication as Head of Asia, where he was part of group management during the trade sale to Lionbridge.
A Comprehensive Guide to Developing Your Machine Translation Road Map
As MT quality improves and use expands, most organizations realize only a small portion of the full business benefits and production efficiencies that MT brings when deployed in a planned and structured manner. This webinar provides a clear, technology agnostic, framework for deploying machine translation beginning with basic utilization and then rapidly maturing through processes, staff roles and key performance indicators.
The presenter explores tools and techniques that will deliver the fastest return on investment and plot these on a roadmap that has measurable performance milestones for the business.
The MT Maturity Model presented in this recording provides an approach to MT maturity with a detailed road map, milestones and metrics that allow organizations to plan their MT deployments and growth in a series of evolutionary steps. An overview of the benefits, efficiencies, and opportunities gained as each milestone is reached is provided as the business that evolves in its technology deployment practices.
- How to progressively plan and deploy MT in a structured manner so as to maximize productivity, margins and returns.
- Provide an understanding of the staffing, process, workflow and technology adjustments required for success.
- How to set KPIs, measure and understand the impact on the business of changes brought about by MT integration.
Listen to the recording below.