Master Project MEDiQ (PFE or master intership, possibility to continue in PhD, 6 months)

Context and Motivation

The Biomedical Data Translator project (NIH/NCATS) seeks to revolutionize disease biology understanding by developing a robust and interpretable machine learning framework that translates complex biomedical data into actionable insights. In a federated and distributed framework, the platform makes recommendations of biological mechanisms involved in diseases to researchers from different disciplines. AI models analyze data across multiple sources, including clinical trials, genomic sequencing, proteomics, metabolomics, etc and provide an explanation of their reasoning that led to the recommendations.

The project aims to improve the recommandations quality by integrating user preferences through reinforcement learning techniques. This project has the potential to identify novel therapeutic targets and uncover underlying biological mechanisms involved in this complex condition.

Objectives

The successful candidate will contribute to building a prototype of Reinforcement Learning (RL) on biological data, specifically focusing on Facioscapulohumeral muscular dystrophy, a muscular rare disease.

The Master’s student will:

• **Conduct a state-of-the-art review of existing RL algorithms employed to improve recommendations
• **Contribute to building a prototype of reinforcement learning to provide new recommandations
• **Validate the approach through testing the quality of the recommendations with the help of a team of experts in the field
• **Participate to the writing of a manuscript on the topic

Candidate Profile

We seek a motivated student with:

• **Proficient in Python programming
• **Knowledge of GitHub or versioning systems
• **Familiarity with Reinforcement Learning (RL) basics, Machine Learning, and/or algorithmics
• **Strong communication skills for collaborating with scientists from diverse backgrounds

Expected Outcomes

The internship will deliver a validated prototype for improving the quality of the answers and has the potential to identify novel therapeutic targets and uncover underlying biological mechanisms involved in this complex condition. The results of this protoype will be further tested in the lab.

What We Offer:

• **Opportunity to work on a cutting-edge project in the field of biomedical data analysis
• **Collaboration with an interdisciplinary team of researchers and clinicians in a very dynamic and cutting-edge informatics laboratory
• **Hands-on experience with RL techniques and their application to real-world problems
• **The chance to contribute to a project that may lead to breakthroughs in understanding and treating rare diseases

If you’re passionate about applying machine learning and computational methods to biomedical research or are seeking a first experience in this field, we’d love to hear from you!

keywords : explanability, reinforment learning, deep learning, machine learning, rare diseases, knowledge graphs

Publications: DOI:10.1109/TKDE.2023.3237741 DOI:10.1145/3331184.333120