À la Une

Soutenance de thèse Athanasios Kyritsis

AK_400.jpg  

M. Athanasios Kyritsis soutiendra en anglais, en vue de l'obtention du grade de docteur en systèmes d'information de la Faculté d'économie et de management (GSEM), sa thèse intitulée:

Enhancing Wellbeing Using Artificial Intelligence Techniques

Date: Jeudi 19 décembre 2019 à 14h00

Lieu: CUI / Battelle bâtiment A, auditoire rez-de-chaussée

 

Membres du jury:

  • M. Gilles FALQUET, Professeur, Président du jury
  • M. Michel DERIAZ, Docteur, GSEM
  • M. Dimitri KONSTANTAS, Professeur, GSEM
  • Mme Katarzyna WAC, Docteure, GSEM
  • Mme Giovanna DI MARZO SERUGENDO, Professeure, Faculté des Sciences de la Société
  • M. Theodoros KOSTOULAS, Professeur, University of Bournemouth
  • M. Yacine BENMANSOUR, État de Genève
 

Résumé:

The landscape of technology is rapidly evolving, and its pace of growth is not slowing down. During the last decades, and especially after the mass adoption of the internet, new technological advances have revolutionized every aspect of human life. We are living in the ubiquitous computing era, where connected devices form the internet of things and produce data faster than we can logically process. With the latest advances in mobile communications and with the widespread use of smartphones and connected sensors, various aspects of wellbeing can be monitored and improved. The goal of this thesis is to propose new algorithms, methodologies, and applications that can be used as components in health and wellbeing systems that support healthy aging, enhance human-machine interactions, and support postoperative rehabilitation with the use of modern connected devices and machine learning techniques. The core research question of the thesis is the following: "How can artificial intelligence techniques ameliorate human wellbeing by using data produced by modern smart devices?" 

In this thesis, we initially investigate how modern technology and applications that support healthy aging can be attractive to be used by older adults. We conducted a research study in order to understand the needs and requirements of older adults, as well as the reasons behind the age technology gap. In this study, we provide useful insights to developers who are building user-centric applications and want to appeal to users of all age groups. 

We are then presenting three components we developed that leverage the use of modern technologies to improve daily living. We are presenting a low-cost, easy to deploy and use, indoor localization system with room-level accuracy. The algorithm we are proposing takes into account signals from radio frequency beacon transmitters like Bluetooth beacons and the room geometry when inferring a position. We continue by presenting a mood and stress detection system that monitors non-invasive sensor data and smartphone usage patterns in order to estimate the psychological state of the users. We are then presenting methods of detecting abnormal behavior from activity and mobile app usage data. Unsupervised anomaly detection techniques were employed for detecting potentially problematic scenarios during the day and for triggering relevant actions. 

In the frame of building context-aware applications, we have also contributed towards developing activity recognition systems using inertial sensors. Initially, we present all the parameters that have an impact on every stage of the development of an activity recognition system. We have analyzed the significance of the parameters in the design, implementation, testing, and evaluation phase of an activity recognition system with an experiment that included several activities to be identified. We put all this acquired knowledge into practice by building an activity recognition system optimized for a specific scenario. We created a gait recognition system targeting people who have undergone lower body orthopedic surgery. We collaborated with the physiotherapist team of Hirslanden Clinique La Colline, an orthopedic clinic in Geneva. We built a system that is meant to be used by the physiotherapists during the rehabilitation phase of a patient, in order to be able to monitor the evolving gait pattern of the patient during everyday life, and not only during the time-limited physiotherapy sessions. 

The thesis contributions include techniques on how machines can learn about different human aspects from data. All the presented components in the thesis can be used to support wellbeing systems that enhance daily life.