Process Innovation Challenge

Track: Advanced Localization Management | AL2 |   Everyone |
Wednesday, July 29, 2020, 11:00am – 12:00pm
Held in: Stream 2
Adam Bittlingmayer - ModelFront 
Justine Chuang - IBM 
Louis Huang - IBM 
Wouter Maagdenberg - TXTOmedia 
Yi-Qun Ren - Micro Focus 
Martin Svestka - Memsource 
John Weisgerber - Transifex 
Andrzej Zydroń - XTM International
Moderator: Dave Ruane
Hosts: Alessandra Binazzi,
Markus Meisl,
Yuka Nakasone

Process Innovation Challenge

The PIC is a platform for innovations and innovators in the localization and translation industry, held during LocWorld conferences. Innovators have a stage to pitch their ideas to peers and experts. The audience will vote for the Process Innovator of the year.

Takeaways: Investors and other interested parties will get to meet innovators in the localization industry and see the latest localization innovations.

AI-enabled Bilingual Terminology Extraction
We made use of the latest advancements in computational linguistic technology to build a robust bilingual terminology extraction feature. This provides highly accurate results in 50 languages and features a user-friendly output format that both professional terminologists and nonprofessional users can use to quickly create a glossary.
Andrzej Zydroń – XTM International

An Innovative Tool to Auto-detect Layout Issues on Webpages
We have created a tool to auto-detect the common layout issues on localized webpages, such as string truncation, overlap and right-to-left rule violation. We are filing a patent for this tool which should be complete in May of this year.
Yi-Qun Ren, Le Peng – Micro Focus

Business Intelligence-powered Post-editing
It is general practice that all machine translated output is post-edited before it is published. But not all content is made equal. Why dedicate time and money on content that receives little-to-no attention? This innovation aims to use valuable post-editing resources where they can bring real business value.
Martin Svestka – Memsource

Bypassing the I18n Problem by Owning It in the Development Platform
We will introduce an open-source self-development kit embedded in app development platforms that scans for new or updated strings, sends them in a universal internationalization (i18n) format (ICU) to a translation management system in “fileless” mode (over the air). The translations are then delivered from a content delivery service at runtime to the development platform for release.
John Weisgerber, Transifex

Speech-to-text Subtitle Translation
Complete sentences in a caption file such as .srt are usually segmented by different time stamps resulting in in poor translation when using machine translation. Our innovation is to use natural language processing technology to combine the segments so they become complete sentences for better translation quality.
Justine Chuang, Louis Huang – IBM

Translation Risk Prediction
We introduce quality estimation as a service— instant risk scores on every translated segment.
Adam Bittlingmayer – ModelFront

Watch Instead of Read
We introduce a technical solution that works on top of content management systems (or standalone) and which transforms structured content into audio and video automatically, immediately creating and maintaining localized how-to videos.
Wouter Maagdenberg – TXTOmedia International BV