2021 Speech Industry Award Winner: AppTek Takes Captioning to New Screens

Article Featured Image

AppTek has a long history of innovation in the areas of artificial intelligence, speech recognition, machine translation, natural language processing and understanding, and text-to-speech technologies, and finally it is getting the worldwide recognition it deserves.

Among AppTek’s many accolades this year, Germany’s Federal Ministry of Health in October awarded its HYKIST speech project to the company, which is based in McLean, Va. The HYKIST project seeks to assist language mediators who work remotely with hospitals throughout Germany, facilitating multilingual communication with non-German-speaking patients through automatic speech recognition and neural machine translation. It combines both technologies with a dialogue system for capturing initial medical history and is integrated into the Triaphon telecommunication platform, which is used to assign language mediators to urgent phone conversations with medical staff in real time.

Then, AppTek’s science director, Hermann Ney, in February received the International Speech Communication Association (ISCA) 2021 Medal for Scientific Achievement for his “pioneering and seminal contributions to data-driven methods for automatic speech recognition and machine translation.”

And in June, AppTek was named a speech recognition provider for CallMiner’s Open Voice Transcription Standard (OVTS) after meeting CallMiner’s rigorous interoperability requirements. By joining OVTS as a founding member, AppTek can integrate with and develop against the CallMiner Eureka platform.

With AppTek’s ASR technology as part of the CallMiner OVTS ecosystem, businesses can now unlock the voice of the customer and agent in more than 30 languages and dialects while relying on AppTek’s AI-enabled speech-to-text technology to uncover actionable insights and business intelligence from 100 percent of voice interactions.

AppTek also began working with Washington’s Gallaudet University, the only university in which all programs and services are specifically designed to accommodate deaf and hard-of-hearing students, to develop a transcription and captioning application for web browsers and videoconferencing platforms.

The new application aims to provide videoconference participants with live closed captions and more control of the user interface. It will draw on AppTek’s ASR platform and also incorporate AI and machine learning. Over time, Gallaudet also intends to incorporate multilingual capabilities using AppTek’s multilingual ASR and neural machine translation technologies.

AppTek also this year partnered with systems integrator and manufacturer PBT EU to integrate its neural machine translation, speech recognition, and intelligent line segmentation technologies directly into PBT EU’s SubtitleNEXT captioning, subtitling, and localization platform.

“AppTek’s scientific expertise and dedication to media and entertainment AI-driven multilingual speech and translation makes them a priority for the SubtitleNEXT platform,” said Ivanka Vassileva, CEO of PBT EU, in a statement at the time. “In addition, innovative value-added features such as [localization] and the ability to implement robust customizations make AppTek the type of partner that will enable us to continue delivering best-in-class services for SubtitleNEXT.”

It’s that same drive and passion that also drove Modzy, an enterprise AI platform and marketplace provider, to partner with AppTek late last year. AppTek was chosen to deploy 16 additional ASR languages and 25 bidirectional neural machine translation-ready languages into Modzy, an enterprise AI platform and marketplace.

AppTek initially provided five ASR languages and two MT languages when the Modzy platform launched in late 2019.

The AppTek languages now available on the Modzy marketplace include Arabic, Belarusian, Bosnian, Bulgarian, Chinese, Croatian, Dari, Dutch, English, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Montenegrin, Pashto, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Spanish, Turkish, Ukrainian and Urdu.

And finally, AppTek in April expanded its Workbench data labeling and annotation platform to include video labeling capabilities for computer vision models.

To create the robust datasets required for high-performing AI models, AppTek employs its proprietary Workbench, a secure, cloud-based data annotation and labeling tool that plugs into any data supply chain and fuses the human-machine relationship to produce high-quality data and generate custom machine learning datasets. The company partnered with IDenTV on this latest offering.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues