Are you eager to investigate new algorithms for the diagnosis of disease by voice?
AI-based voice diagnostics analyse spoken audio using machine learning methods. Responsible application of AI-based voice diagnostics holds great promise in healthcare because of its usefulness in diagnosing and monitoring disease. This project develops AI-based voice diagnostics for the most common neurodegenerative diseases, Parkinson's and Alzheimer's, using speech analysis. This project closely collaborates with medical experts in the area of neurodegenerative disease to understand how to responsibly collect data necessary to train AI-based voice diagnostic models. It also addresses the challenge of fine-tuning foundation models for classification under limited data conditions. Finally, it tackles a key shortcoming of current models trained for the detection of neurodegenerative disease: the lack of explainability of these models makes it difficult to determine the reasons behind correct and incorrect predictions.
The project will leverage recent ideas from data-centric AI, explainable AI and data minimisation. As a PhD candidate, you will make a creative contribution in putting these ideas to work for supporting the protection of spoken audio recordings in the setting of medical diagnosis. You will work in a multidisciplinary group consisting of PhD candidates, postdoctoral researchers and professors in different domains such as linguistics, language and speech technology, speech recognition (ASR), NLP, data privacy and responsible AI.
You will be responsible for 1) creating speech data collection specifications and risk assessment in collaboration with stakeholders from the medical domains, 2) developing and testing algorithms that are able to diagnose disease on the basis of speech data, making use of the principles of responsible AI, and 3) designing and validating novel techniques for explaining the results of algorithms, making them more useful in a clinical setting. You will be expected to publish your results in leading conferences and journals. The position may also include other duties such as supervision of Bachelor’s and Master’s students.
Fixed-term contract: You will be employed for an initial period of 18 months, after which your performance will be evaluated. If the evaluation is positive, the contract will be extended by 2.5 years (4 year contract).
Houtlaan 4, 6525 XZ, Nijmegen
This application process is managed by the employer (Radboud University). Please contact the employer for questions regarding your application.
Please contact the employer for questions regarding your application.
Tip: save this job as favorite in your AcademicTransfer account. This gives you an immediate overview and makes it easy to find the job later on. No account yet? Create it now and take advantage of other useful functionalities too!
Make sure to apply no later than 17 Mar 2024 23:59 (Europe/Amsterdam).
We like to make it easy for you, sign in for these and other useful features: