Exhibitors & Products
Events & Speakers

– "Lecture Translator" system developed by Karlsruhe Institute of Technology (KIT) provides automated simultaneous translations of human speech

Karlsruhe/Hannover. Self-learning systems will never replace humans, but they can certainly help and support them – for instance, with global communication and intercultural dialogue. Germany's Karlsruhe Institute of Technology (KIT) is a world leader in the field of self-learning systems for human language understanding. Its Interactive Systems Lab (ISL), part of the Institute for Anthropomatics and Robotics (IAR), has developed Lecture Translator, a digital system for automated simultaneous translation that can be installed in lecture auditoriums and used in mobile applications. The translations are output as transcripts, which users can follow on screen. Lecture Translator uses spoken language translation (SLT), a blend of technologies that combines automatic speech recognition (ASR) with machine translation (MT) plus various auxiliary functions that structure the text, deal with challenges like punctuation and German's proclivity for large compound words, record the lecture and display the translation result on screen.

Translating lectures

KIT installed a version of Lecture Translator as a permanent system in various lecture auditoriums as long ago as 2012 as part of its efforts to help foreign students follow German-language lectures dealing with complex scientific and technical subject matter. The spoken German is simultaneously translated and the results displayed in text form during the lecture. "Lecture Translator uses a cloud-based service infrastructure," explains Alexander Waibel, Professor of Computer Science at KIT and IAR director. "The spoken language is captured by a local client and forwarded to the cloud infrastructure, where a service manages the flow of data through the ASR, MT and other components." The English-language output is displayed online at https://lecture-translator.kit.edu during the lectures. The system also archives the lecture transcripts so that students can go back to them and run text searches. Other institutions of higher learning have expressed interest in the Lecture Translator system.

Applications beyond education

The system also has applications outside of teaching and education. Its speech recognition component can also be used to generate real-time transcripts of speeches, in the form of subtitles, for example. In this way, Lecture Translator can, for instance be used to enhance access to spoken content for people with hearing impairments. It can also be used in this way in Internet and TV broadcasts.

Cooperation with industry

Software-driven communication tools deliver efficiency and transparency benefits across many industries. The Interactive Systems Lab at KIT has been developing highly promising applications of its automatic simultaneous translation technology in cooperation with partners in several industries. For example, in a joint project with SAP, ISL has adapted its Lecture Translator system for the massive open online courses (MOOCs) that SAP offers as part of its openSAP free online training program. The result of this project was showcased in 2017 at SAP's own in-house tradeshow SAPPHIRE NOW in Orlando, Florida. ISL and SAP are currently working on plans for another joint project, which will focus mainly on automatic speech recognition.

ISL partnered with training provider alfatraining Bildungszentrum to showcase the potential applications of Lecture Translator in video conferencing software at the education trade show LEARNTEC, in Karlsruhe. KIT and alfatraining are also taking part, along with several other partners, in ELITR, an EU research project on the use of translation software in video conferences.

KIT will be running a live demonstration of its Lecture Translator system at Stand B16 in Hall 2, home of the Research & Technology show, at HANNOVER MESSE 2019 (1 to 5 April). KIT will also have a presence at HANNOVER MESSE's Integrated Energy show (Stand L51, Hall 27).

Your contact for further information:
Karlsruhe Institute of Technology (KIT)
Monika Landgraf
Tel: +49 721 608-21150