Lectures on Models of Human Behaviour

Currently computers are being moved out of the desktop setting and into our daily lives. 
The miniaturization of computers and the introduction of mobile phones has made computers 
more embedded in our daily lives, so one of the next great challenges is the new interface 
between human and computers. Also novel computer systems are enriched with all sorts of 
sensors to measure properties of their environment, allowing computers to be even more 
aware of their surroundings. This novel computing paradigm is known under a variety of names, 
namely: pervasive computing, ubiquitous computing, ambient intelligence and context awareness. 
This new trend in computing allows for many different branches of applications that focus on 
various aspects of our lives. Applications for these systems rely on different kinds of contextual 
information, such as location, identity of the user, and the activity a user is performing. 
Recognizing human activities from smart home sensor data allows many applications, in areas 
such as intelligent environments and healthcare.

Well known journals as IEEE Pervasive Computing, IEEE Transactions on Information Technology 
in Biomedicine, Personal and Ubiquitous Computing dedicate issues to topics related with sensor 
networks and behaviour modelling. These lectures will highlight how computers can create models 
of human behaviour from sensor readings. Current state of research will be presented, focusing 
both on the theoretical machine learning techniques to infer the behaviour of humans and on the 
practical issues such as the trade-offs between the informativeness of sensors and privacy issues, etc.