Patrick Reignier

Learning Situation Models in a Smart Home

This paper addresses the problem of learning situation models for providing context-aware services. Context for modeling human behavior in a smart environment is represented by a situation model describing environment, users, and their activities. A framework for acquiring and evolving different layers of a situation model in a smart environment is proposed. Different learning methods are presented as part of this framework: role detection per entity, unsupervised extraction of situations from multimodal data, supervised learning of situation representations, and evolution of a predefined situation model with feedback. The situation model serves as frame and support for the different methods, permitting to stay in an intuitive declarative framework. The proposed methods have been integrated into a whole system for smart home environment. The implementation is detailed, and two evaluations are conducted in the smart home environment. The obtained results validate the proposed approach.

Reinforcement Learning of User Preferences for a Ubiquitous Personal Assistant

New technologies bring a multiplicity of new possibilities for users to work with computers. Not only are spaces more and more equipped with stationary computers or notebooks, but more and more users carry mobile devices with them (smart-phones, personal digital assistants, etc.). Ubiquitous computing aims at creating smart environments where devices are dynamically linked in order to provide new services to users and new human-machine interaction possibilities. The most profound technologies are those that disappear. They weave themselves into the fabric of everyday life until they are indistinguishable from it (Weiser, 1991). This network of devices must perceive the context in order to understand and anticipate the users needs. Devices should be able to execute actions that help the user to fulfill his goal or that simply accommodate him. Actions depend on the users context and, in particular, on the situation within the context. The objective of this work is to construct automatically a context model by applying reinforcement learning techniques. Rewards are given by the user when expressing his degree of satisfaction towards actions proposed by the system. A default context model is used from the beginning in order to have a consistent initial behavior. This model is then adapted to each particular user in a way that maximizes the users satisfaction towards the systems actions.