Philips Marie Curie Fellowship HEART ESR 1: Health Behaviour Analytics on Heterogeneous Data in Eindhoven, Netherlands
PhD student for the Marie Curie project HEART. Work will be data analytics for Personal Health.
ESR1 - Health Behaviour Analytics on Heterogeneous Data
HIRING INSTITUTION: Philips Electronics Nederland B.V. (the Netherlands)
PHD ENROLLMENT: PhD position in Computer Science at the Arenberg Doctoral School KU Leuven (Belgium, http://set.kuleuven.be/phd)
You will work in the Personal Health department of Philips Research Europe. The department develops solutions that empower people to manage their health and support professionals in providing better care. A strong focus of the research is on measuring and monitoring people’s health status and habits in their own home and to provide motivating feedback to come to a healthier lifestyle. Application domains that the department is focusing on are cardiovascular, sleep, and elderly care, where the group is closely collaborating with the business units of Personal Health Solutions, Sleep and Respiratory Care and Population Health Management. The department has various clinical collaborations across Europe.
Human activity recognition and vital sign monitoring play a significant role in tailoring personal health and behaviour change coaching solutions to each individual. With the advent of wearable sensors and devices, a unique opportunity for healthcare is emerging where it becomes possible to monitor the health parameters and health related parameters for continuously for long periods, outside the lab or doctor’s office setting, and for large numbers of people. This represents a unique opportunity to gain insight into how behavior and lifestyle influences the health of people and in how changing of lifestyle can improve the health of people.
The following objectives are proposed:
1) Activity recognition and detection of critical situations: Recent wearable devices measure multimodal data streams. Different physiological properties are measured such as physical activity, galvanic skin response and heart-rate. Other data include in-home Internet-of-Thing (IoT) sensors and device data, smart phone sensor data and interaction data. The objective is to combine these heterogeneous data streams from wearable and IoT sensors to recognize health-related activities of the user and physiological parameters to extract actionable insights for example detection of critical situations as e.g. epilepsy seizures, unusual heart rhythms or unusual behavior and activities.
2) Relation between activities and health parameters: Aggregating parameters as activity levels, vital signs and context will allow to create a person specific health profile. This health profile consists of different levels: a) health parameters (heart rate, blood pressure, weight, mental wellbeing); b) behavior and activity related parameter, such as the amount of sleep, physical activity, amount of stress full activities or relaxing activities, amount of work-related activities, food intake, social activities, etc. The objective is to use data analytics and machine learning to find person-dependent patterns in the link between behaviors and health parameter.
3) Data-acquisition: ESR1 will support the data-acquisition of a multimodal database consisting of motion, & vital sign modalities. An efficient approach to come to a high-quality labelled set is investigated.
These objectives will likely involve analysis of time series data with new advanced methods, such as, deep learning. A key challenge is to combine common knowledge with pattern recognition methods. 9
Algorithms for the recognition of complex activities based on multimodal and heterogeneous data.
A multimodal data-driven health-profile classifier based on vital sign and accelerometer data.
Quality runtime code for the use of these such models, calibrated on preliminary and historical data.
INDICATIVE PLANNED SECONDMENTS - Institution, place and timing expressed in contract month (M)
University of Macerata (Macerata, Italy) - M4
KU Leuven (Leuven, Belgium) - M7-8 ; 9-11
Fudan University (Shanghai, China) - M15-19
University of Macerata (Macerata, Italy) - M25
KU Leuven (Leuven, Belgium) - M34-36
Further analysis might be required, based on the development of the research project.
SUPERVISORS : Dietwig Lowet (firstname.lastname@example.org), Bart Vanrumste (email@example.com), Stijn Luca (firstname.lastname@example.org – www.kuleuven.be/advise )
ADDITIONAL ESSENTIAL REQUIREMENTS :
Master degree with distinction (cum laude) in computer science (or equivalent).
DESIRABLE REQUIREMENTS: a clear interest in and knowledge of machine learning, e.g. deep learning. Expertise in data analytics/classification and affinity for computational modeling. Good programming skills are desired. Knowledge of either Matlab/Pyhton/R is required.
Download the complete Call for application via this link:
Please apply via