PhD position in Neural Architectures for Speech Technology

In the context of a Swiss NSF grant, we seek a PhD student to work in the general area of neural architectures for speech technology.

The particular focus of this thesis will be on physiological models of speech production and perception, and how such models can be integrated into deep learning solutions.

The research will build on work done over the past few years at Idiap on creating tools for physiologically plausible modelling of speech. The current "toolbox" contains rudimentary muscle models and means to drive these using conventional (deep) neural networks. More recently, the focus has been on theoretical underpinnings via rigorous Bayesian techniques.

We expect the work will commence with a study of current "neural" vocoders, focusing on those based on physiological models and on normalising flow. Although the project remit is quite open, a significant research thread will be to factorise these models into physiological and probabilistic components; this will be with a focus on identifying how they may be controlled by external agents such as dialogue managers. Another possible thread is to examine these models in the context of speech recognition. In doing this, we hope not only to enable the next generation of expressive speech recognition and synthesis, but also to make inference about the underlying physiological mechanisms of speech production and perception.

The ideal Ph.D student should have a master (or equivalent) degree in engineering, physics, applied mathematics or similar. Graduates of neuroscience programs would also be well qualified. S/he should have a good background in mathematics, statistics, and programming (C/C++, Python, scripting languages). In order to balance the group, we especially encourage female applicants. However, all applications will be judged on merit.

The student will be registered at EPFL and located at Idiap. The position is available from spring 2021, fully funded for four years. Selection will commence in the new year and continue until the position is filled.

Interested candidates are invited to submit a cover letter, a detailed CV, and the names of three references through the Idiap online recruitment system:

PhD position in Neural Architectures for Speech Technology