Postdoctoral Researcher (f/m/d) in the area of Machine Learning and Speech Signal Processing

At the department of Computer Science and Mathematics, the Cognitive Systems Lab (Prof. Tanja Schultz) offers a full-time position (asap for a maximum duration of 24 months) of a

Postdoctoral Researcher (f/m/d)
TV-L 13 (100%)

in the area of Machine Learning and Speech Signal Processing under the MyVoice project, funded by DFG.

Project description:

The aim of the research project is to implement and verify a silent speech interface which enables users to communicate silently while being understood communication partners. For this purpose, electromyography signals obtained from the articulatory muscles are processed by machine learning methods in order to synthesize the otherwise missing acoustic signal. Besides a natural speech output, the project focuses on a real-time synthesis, such that the application of user-in-the-loop methods is possible. Users can alter their silent articulation based on the system output while the system is continuously adapted using machine learning methods such as active or few-shot learning. The goal of the project is that silent speech interfaces should be usable in everyday communication scenarios after a short adaptation phase.

The research is carried out in the Cognitive Systems Lab at the Universit t Bremen in collaboration with the Istituto Dalle Molle di Studi sull'Intelligenza Artificiale in Switzerland. The tasks include the development of protocols and software for data recordings, the development of training and adaptation procedures for machine learning models, and the conduction of verification experiments for the developed Silent Speech Interface.

Specific requirements:

• Excellent doctoral degree in Computer Science or closely related fields
• Solid research experiences in machine learning and signal processing, particularly with speech-related acoustic and/or muscle activity signals
• Publication record in the field of speech processing, preferably with publications in relevant international conferences, such as Interspeech, ICASSP, ASRU
• Excellent programming skills and experience in deep learning libraries, such as PyTorch, TensorFlow, or JAX
• Interested to conduct lab experiments with volunteer subjects
• Excellent communication skills and enthusiasm to work in a team
• Very good knowledge of English language

Application documents should include a letter of motivation, a CV, certificates, the applicant s research and technical background as they relate to the position, and two reference letters. Please send your application in English, quoting the reference number A282/22 by October 20th, 2022 to:

Prof. Dr.-Ing. Tanja Schultz
Cognitive Systems Lab, Universit t Bremen
Enrique-Schmidt-Stra e 5
D-28359 Bremen, Germany

or by e-mail in electronic form (please as one PDF file) to the following address: scheck@uni-bremen.de

We kindly ask you to send us only copies (no portfolios) of your application documents, as we cannot return them.

Shortlisted candidates will be contacted and invited to an interview via zoom.

The University of Bremen aims at increasing the number of women in science and therefore explicitly encourages applications from female candidates. In the case of equal personal aptitudes and qualification, priority will be given to disabled persons. Applicants with a migration background are welcome.

For further enquiries please contact

Prof. Dr. Tanja Schultz
Cognitive Systems Lab, Universit t Bremen
Enrique-Schmidt-Stra e 5
D-28359 Bremen, Germany
or by e-mail to: scheck@uni-bremen.de