Using deep bidirectional recurrent neural networks for prosodic-target prediction in a unit-selection text-to-speech system
- INTERSPEECH 2015
Ron Hoory is the manager of the Speech Technologies group at IBM Haifa Research Lab. His expertise and research interests are in the area of speech processing, including speech synthesis, speech recognition and speaker recognition. He received his B.Sc. and M.Sc. degrees in Electrical Engineering from the Technion, Israel Institute of Technology, Haifa, Israel, in 1990 and 1993, respectively. He joined IBM Haifa Research Lab in 1993 and led research and development activities on embedded concatenative text-to-speech, distributed speech recognition, Hebrew speech recognition and very low bit rate speech coding. In 2006 he was appointed group manager. During 2009-2014 he led the the research work on text-to-speech and speaker verification within the JDA with Nuance. In 2014 he was appointed Senior Technical Staff Member (STSM). In 2015-2016 he led the research & development of the new Watson TTS Service, and since then he continued to lead the global TTS research and transformation of the TTS service to a natural sounding, expressive and conversational TTS based on deep nerural networks. Since 2019 he has been the spoken customer care subtheme co-lead.