Synopsis
Like many areas of information processing, automatic speech processing has been revolutionized by deep learning. In this presentation, I will address various problems in speech generation through paradigms such as text-to-speech (TTS), voice enhancement, and voice conversion. I will provide a state-of-the-art overview of these technologies and their applications (voicebots, deepfakes, assistive technologies, etc.). Finally, I will present some recent work carried out at the MIAI Institute aimed at improving the expressiveness, responsiveness, and controllability of speech generation systems.