RTLC 2012 - Workshop on Recommender Technologies for Lifestyle Change 2012

In today's society, particularly in the affluent society, lifestyle is influenced by technology and the abundance of financial resources. For instance, a large variety of computer games are heavily used and people often travels by individualized transportation means, such as car, just for fun. The idea that technique and money can buy anything spreads also to health management: people believe that nowadays medical knowledge can be immediately applicable in case of illness, as technical knowledge can be used for repairing a broken car.
This results in lifestyles that do not care about the negative long-terms effects on the environment, but also on each individual person. The most prominent example of this is represented by various types of chronic illnesses in developed countries that result from poor lifestyle choices.
In this context, the aim of this workshop is to explore possibilities for recommender systems to support users in taking decisions related to various aspects of their lifestyle; we call them Lifestyle Change Recommender Systems (LSCRS). There are three main challenges for LSCRSs: firstly, such systems have to assess the user's context for delivering such recommendations. Secondly, in order to promote any change in the lifestyle of the user, they have to recommend a tailored sequence of items, mostly actions, taking into account the dependencies between items and the effects of each item recommendation. Thirdly, the systems have to be defined in a way that favors the user's continuous attention, and allows to explain the reasons for the change in the user's future behaviour, and to communicate the changes already effectuated.
Solutions to the first challenge amount to formalize knowledge in various domains from two perspectives: on the one hand, from a user's point of view, options for actions in a certain situation have to ready to be rated in order to assess user preferences in the domain. On the other hand, domain knowledge is necessary to valuate each option from an expert's point of view. A good recommendation by a LSCRS both satifies user preferences and is compatible with expert knowledge. This property of a LSCRS is very challenging in terms of its algorithmic implementation and the acquisition and formalization of the required domain knowledge.
Beyond that, in order to assess up-to-date information about context, recommender systems need to have access to relevant sensor data and allow users to provide input and feedback about context and the recommended options. i
The second challenge requires researchers to develop algorithms that can look several steps ahead in the future in order to recommend sequences of items in which for the recommendation of each item, the item recommended before is taken into account. In this challenge recommender systems meet planning systems.
For the third challenge, it is important to observe that while many people are aware of these problems, they lack sufficient knowledge an information that would help them to take preventive measures that are both suited for their personal needs and fit the rhythm of their every day life. Even if people know that they should do more sport and eat meals with healthy ingredients, they are often unsure regarding how much is enough, the type of exercise to do, or which ingredients, in which combinations and quantities are appropriate. Similar arguments hold for other every day activities.
Hence, in order to provide an effective support to lifestyle change, recommender systems need to provide communicative capabilities, e.g, with multi-modal dialogue systems. Recommender technologies must initiate a feedback-change-loop that could contribute to lowering the risks of severe illnesses for many individual users and to improving the overall environmental situation.
In the light of these challenges, submissions are solicited that address the following topics:
Surveys of lifestyle related activities and technological approaches to monitoring them
Context modeling for activity recommendations
Formal models of sensor data for monitoring every day activities
User models for every day life recommendations that provide user-tailored content
Motivational models for lifestyle, every day activities, and environmental responsibility
Recommendations of sequences of items (e.g. physical exercises for a whole week, planning meals for a month)
Measures of the effectiveness for lifestyle change recommender systems
Approaches to combine sensor data and interactive user input in LSCRSs
Strategies to cement behavourial change
Strategies for situation- and user-aware presentation of recommendations
Persuasive technologies for interaction with and/or among users on their personal situation, their habits, and their options to change their lifestyle
Recommendation of activities for leisure time and lifestyle
Recommendation of information sources (e.g. forum entries, blogs) for LSCRSs
The workshop will be an opportunity for discussing and promoting ideas on how intelligent information systems can provide users with up-to-date information in every day life situations to make choices for every day life activities that establish a sustainable compromise between quality of life, individuality, and fun.