Internship in Health Data Science/Bioinformatics/Biostatistics - Vocovid Project - GF0921
1A-B, rue Thomas Edison, Strassen L-1445, Luxembourg
The evolution of voice technology, audio signal analysis, and natural language processing/understanding methods have opened the way to numerous applications of voice tracking and analysis in the field of healthcare. Voice is a rich medium, in which a lot of information can be found, providing that we know how to isolate and make use of it. A vocal biomarker can be defined as a signature, a feature, or a combination of features from the audio signal of the voice that is associated with a clinical outcome and can be used to monitor patients, diagnose a condition or grade the severity or the stages of a disease or for drug development. Covid-19 is a respiratory condition, affecting breathing and voice, and causing, among other symptoms, dry cough, sore throat, excessively breathy voice and typical breathing patterns. These are all symptoms that can make patients’ voices distinctive, creating recognizable voice signatures and enabling the training of algorithms to predict the presence of a SARS-COV-2 infection or as a tool to grade the severity of the disease. In our context, voice is a promising data source to identify vocal biomarkers associated with Covid-19-related outcomes such as respiratory symptoms, fatigue or stress. Developing Covid-19 vocal biomarkers would help to offer a better remote monitoring of people with Covid-19 based on digital technologies.
Our team aims at identifying vocal biomarkers of Covid-19 related symptoms, based on audio recording from the large prospective Predi-COVID cohort study.
Training and research environment
The “Deep Digital Phenotyping Research Unit” develops a research activity within the Department of Population Health at the frontier between digital epidemiology, digital health and data science, to leverage real-world data to improve population health.
The Master student will directly contribute to the identification of vocal biomarkers for Covid-19 monitoring. He/she will lead a project on the analysis of a large annotated audio dataset from the Predi-COVID cohort study to decompose the audio signals and then apply AI-based algorithms and statistical approaches to identify features associated with Covid-19 related outcomes. He/She will have to take charge of the literature review, the pre-processing of the data, the data analysis, and the preparation of a scientific article. He/she will be supervised by Dr. G. Fagherazzi, Director of the Department of Population Health, and Aurélie Fischer, Ph.D. student in the DDP lab, in association with experts in audio signals and artificial intelligence methods.
This internship position may lead to a Ph.D. opportunity in digital epidemiology.
KEY SKILLS, EXPERIENCE AND QUALIFICATIONS
● Master level in Data Science, Bioinformatics, Biostatistics or equivalent required.
● Flexibility, adaptability, autonomy
● Ability to work with different profiles and people from different experiences and cultural backgrounds
● Language skills: Fluency in French and English is an asset, with written skills in both languages. Any other language in use in Luxembourg would also be an asset.
Guy Fagherazzi, PhD
1A-B, rue Thomas Edison
What we offer and conditions
- Students will have the opportunity to work in an interactive and international scientific environment, attend conferences by eminent scientists from abroad, and present their own work during lab meetings.
- They will receive training in digital epidemiology research and data science and will have the opportunity to gain skills in data analysis, AI methods applied to audio data.
- A compensation of 660,58€/month will be paid.
- Applicants should be enrolled in a Master program and the internship should be a mandatory part of the diploma.
- Students from abroad can apply for an Erasmus grant.
- Contract type: 6-month fixed-term contract
- Work hours: Full time- 40h/week
- Location: Luxembourg
- Start date: According to your availabilities
- Ref: JF/INTVOCOVID0921/GF/DDP