Master’s Internship Project Privacy-Preserving Detection of Long-Term Behavioral Drift in Older Adults Using a Hybrid Learning Approach

2 Internship Tasks This project aims to investigate the benefits of a hybrid learning approach based on split learning for detecting long-term behavioral drift. The project will be organized around the following tasks: • Conduct a comprehensive literature review of existing research on long-term behavioral drift detection. The objective will be to identify existing approaches to long-term behavioral modeling, the types of data used, and the datasets that could be used to validate our proposed methods. • Conduct a comprehensive literature review of existing split learning architectures and the different possibilities for partitioning the learning process. • Investigate the suitability of different learning architectures for long-term behavioral drift detection, considering both computational performance and patient privacy. • Implement centralized and distributed simulations of deep learning approaches on relevant datasets to detect behavioral drift. • Evaluate the different models and compare them with a centralized baseline. Based on the selected evaluation metrics, the proposed hybrid learning approaches will be assessed and compared with a fully centralized solution. 3 Eligibility Requirements • Applicants must have completed, or be currently enrolled in, the second year of a Master’s degree program. • Knowledge of privacy and/or machine learning will be considered an asset. 4 Application Procedure • Applicants must submit their application by email, including the required documents (see Section 6). • The application deadline is November 8, 2026. • An interview to assess the applicant’s motivation may be arranged in late November. 5 Internship Details • The internship is expected to start between February and March 2027. • Duration: 6 months. • The internship is paid and involves 35 working hours per week. • The internship may be carried out either at INSA Lyon (LIRIS) or at INSA Haut-de-France (LAMIH) in Valenciennes. 2 6 Required Application Documents • Curriculum vitae (CV). • Master’s Year 1 (M1) and Master’s Year 2 (M2) academic transcripts (including Semester 1 results if the academic year is still in progress). • Cover letter explaining the applicant’s interest in the project and highlighting relevant skills and experience. 7 Acknowledgement This work is part of an OpenLabs project aimed at strengthening scientific and technical cooperation between laboratories across the INSA network.