The course will include topics in mobile sensing, mobile on-device sensor data processing and applications thereof.
1. Introduction to mobile sensing platforms and the applications of mobile sensing;
2. Foundations of sensor sampling and properties of sensing hardware;
3. Machine learning for mobile sensing, data processing pipeline (from raw data to high-level inferences);
4. Deep learning for sensor data processing;
5. Machine learning on edge devices. Machine learning model compression and optimisation techniques;
6. Federated learning in theory and practice using FLOWER framework on Android;
7. Location sampling, trade-off between accuracy and energy usage;
8. Location tracking and prediction, Markov chains for trajectory modeling;
9. Physiological signal sampling and processing; photoplethysmogram, skin temperature, electrodermal activity sensing;
10. Voice sensing and processing; emotion and stress inference from voice data;
11. Attention inference and notification management in mobile computing;
12. Sensing for health and wellbeing; digital behaviour change intervantions;
13. Mobile and IoT sensing for security;
Practical part of the course consists of programming labs, while the students also complete a semester-long sensing and data processing programming project.
