DeepVoice is the Parisian event dedicated to voice and artificial intelligence with the aim of bringing together technical expertise and business/entrepreneurial innovation.

This 2-day event will combine a keynote presentation, two practical workshops, a thematic round table and networking moments in order to foster exchange and collaboration amongst all participants. 

June 23

Morning - 10h

Keynote - Expressive Neural Text-to-Speech - Thomas Drugman (Amazon)

Recent advances in Text-to-Speech (TTS) based on Deep Learning have made artificial voices almost indistinguishable from human speech in short neutral content. A substantial gap still exists however, when it comes to how AI-based systems are able to render emotions. This talk will focus on our efforts, in Amazon TTS, to close this gap by making our generative models more expressive and human-like. We will see how this technological progress improved products such as Amazon Alexa or AWS Polly.

Afternoon - 14h

Workshop #1 - Neural speech synthesis - Nicolas Obin (Ircam)

June 24

Morning - 10h 

Roundtable - Assistants, emo/co-bots and virtual reality: how will the simulation of emotions modify our emotional relationship with machines and our immersion in an artificial social world? 

Susana Sánchez Restrepo (SoftBank Robotics), Nadia Guerouaou (CHRU Lille), Catherine Pelachaud (Sorbonne University), Jean-Julien Aucouturier (CNRS), Serge Tisseron (Université de Paris, Académie des technologies) / Nicolas Obin (moderator) 

 

Afternoon - 14h 

Workshop #2 - Neural vocoder - Axel Röbel (Ircam)