Unbabel Discusses the Future of Its Ultimate Demo of Human-Verified Translations for Voice Instant Messaging at the World’s Largest Conference in Speech Processing
Unbabel presents its new demo of Human-Verified Translations for Voice Instant Messaging at the 20th Annual Conference of the International Speech Communication Association INTERSPEECH 2019, held in Graz, Austria, from September 15-19, 2019. Interspeech is the world‘s largest and most comprehensive conference on the science and technology of spoken language processing.
Being the ultimate demo of Unbabel, Unbabel Talk is a speech-to-speech translation application that provides human-verified translations for voice instant messaging (IM) in multilingual scenarios. It combines Unbabel’s translation pipeline with state-of-the-art automatic speech recognition (ASR) and text-to-speech (TTS) models, to send a voice message in a language of choice through popular messaging platforms. High-quality translations are ensured through Unbabel’s own quality estimation (QE) tool and/or through Unbabel’s community of translators.
Presented to over 2000 leaders, experts, and students in speech processing, Unbabel will discuss the future of its latest demo, that is currently divided into two phases:
- The current version: which synthesizes audio that can be delivered with male or female standard voices;
- Forthcoming versions: with features in development, such as voice morphing and transcription correction through Unbabel’s community.
Unbabel Talk will soon be available on the App Store, enabling the public to exchange multilingual voice messages on IM platforms such as WhatsApp, Facebook Messenger, and others.