PhD position “Unsupervised/semi-supervised learning algorithms for speech enhancement and source localization”

The Signal Processing Division and the Collaborative Research Centre Hearing Acoustics at the University of Oldenburg in Germany are seeking to fill the position of a

Research Scientist (PhD Student) - “Unsupervised/semi-supervised learning algorithms for speech enhancement and source localization”

The position is available from 01.08.2023 for 3 years, with salary according to TV-L E13 (75%), corresponding to about 3.200 € per month before taxes (exact amount depending on experience and qualifications).  

The main activities of the Signal Processing Division (https://uol.de/en/mediphysics-acoustics/sigproc) centre around signal processing for acoustical and biomedical applications, with a focus on hearing aids and speech communication devices. More specifically, research topics in the areas of microphone array processing, speech enhancement and acoustic scene analysis are addressed, using a combination of model-based statistical signal processing techniques and data-driven machine learning methods. The Signal Processing Division has access to excellent high-performance computing facilities, measurement equipment and labs, e.g., a unique lab with variable acoustics.   

The Collaborative Research Centre Hearing Acoustics (https://uol.de/en/sfb-1330-hearing-acoustics) aims at a fundamentally better quantitative understanding of the principles underlying the processing of complex auditory and audio-visual scenes, the implementation of this knowledge in algorithms for perceptual enhancement of acoustic communication, and the evaluation of these algorithms for different applications. The successful candidate is expected to investigate unsupervised/semi-supervised learning algorithms for speech enhancement and source localization within a hybrid computational acoustic scene analysis (CASA) framework. Using this CASA framework, we aim at leveraging the potential of recent machine learning methods while maintaining the interpretability of conventional signal processing modules through high-level interpretable latent variables.

Responsibilities/Tasks

• carry out research on acoustical signal processing algorithms for speech enhancement and source localization, involving algorithm design, implementation, and experimental validation;
• write scientific papers for international conferences and journals;
• actively participate in the research meetings and seminars at the Department of Medical Physics and Acoustics

Profile

• Candidates are required to have an academic university degree (Master or equivalent) in electrical engineering, engineering physics, hearing technology and audiology or a related discipline, excellent grades and a solid scientific background in at least two of the following fields: speech and audio signal processing, machine learning, acoustics.
• Familiarity with scientific tools and programming languages (e.g., python) as well as excellent English language skills (both oral and written) are required.
• For the envisaged research project, experience with unsupervised/semi-supervised learning methods and acoustical signal processing algorithms is beneficial.

For applicants outside of the European Union it is highly recommended to check if your academic university degree is equivalent to a German higher education qualification. Please consult the website of the Central Office for Foreign Education (https://www.kmk.org/zab/central-office-for-foreign-education.html) for more information and to apply for a statement of comparability.

The University of Oldenburg is dedicated to increasing the percentage of women in science. Therefore, equally qualified female candidates will be given preference. Applicants with disabilities will be preferentially considered in case of equal qualification.

To apply for this position, please send your application (ref. SP232) including a letter of motivation with a statement of skills and research interests (max. 1 page), curriculum vitae, and a copy of the university diplomas and transcripts to simon.doclo@uol.de. The application deadline is 21.04.2023.