Internship - Multimodal Patient Modelling and Knowledge Graphs for Personalized CPAP Follow Up
Internship period: 02-07 / 2027
Required profil: M2/Engineer Computer science
Internship Context
This internship contributes to a doctoral research collaboration between Linde Homecare France and Université Lumière Lyon 2, through the DISP Laboratory, focused on personalized follow-up for patients receiving Continuous Positive Airway Pressure (CPAP) therapy. The research aims to connect early detection of treatment difficulties, evidence-based assessment, personalized interventions and patient responses to support treatment adherence and long-term self-management. A central research challenge is preserving and exploiting knowledge from heterogeneous patient information collected over time, including CPAP monitoring, questionnaires, care history and intervention feedback. Beyond integrating these data, the framework must maintain the relationships between patient context, available evidence, professional decisions, interventions actually delivered and observed outcomes, while accounting for missing information, temporal changes and uncertainty. To address this challenge, the research investigates knowledge graphs as a means of representing, updating, analyzing and retrieving contextualized treatment experiences. The objective is to determine whether knowledge retained from previous patient episodes can support future follow-up decisions, while evaluating the added value of graph-based representation compared with simpler data structures. This internship focuses on designing and experimentally evaluating a temporal knowledge-graph prototype using available patient records and existing assessment outputs. The work includes knowledge modelling, graph construction and updating, contextual experience retrieval, and reproducible evaluation of representation correctness, traceability and retrieval relevance. The results will contribute to the doctoral research on knowledge continuity and informed reuse for personalized CPAP homecare.
Internship Mission Description
The intern will design, implement and experimentally evaluate a temporal knowledge graph for representing and retrieving contextualized CPAP treatment experiences. The prototype will connect patient evidence, professional decisions, interventions and observed responses while preserving their temporal relationships, provenance and uncertainty. The objective is to investigate how retained treatment experiences can support future personalized follow-up and evaluate the added value of graph-based representation.
Additional details about the Internship Mission