Preliminary schedule online
Invited talks by Konstantin Schekotihin and Bruno Zanuttini
Workshop will be held jointly with MPREF workshop CEUR-WS proceedings available
Finding the information or product that matches your needs and preferences on the Web can be challenging. Recommender Systems (RS) have proved to be helpful tools for various information seeking and filtering tasks and are nowadays ubiquitous on the Web. The most prominent classes of such systems are based on the detection of preference patterns in larger user communities (collaborative filtering) or on the automatic construction of content-based preference profiles for individual users (content-based filtering).
For some application domains in e-commerce, relying on such pattern or model learning approaches alone can be insufficient because explicit and detailed short term interest or preference profiles are required to determine the most suitable product. This is for example the case when (a) the products themselves are configurable and can be tailored to the specific needs of a user or (b) the selection of the right product is based on user-specified explicit constraints, e.g., when purchasing a digital camera or smartphone. Going beyond collaborative and content-based techniques, such knowledge-based approaches therefore often require more complex user interfaces which support an interactive preference elicitation or explanations as well as explicitly encoded domain knowledge.
The crucial issues in this area for example include:
- efficiently reasoning and representing the preferences of the user (allowing the specification of "complex" preferences which can involve well-refined tradeoffs between different criteria),
- learning preferences during the interaction,
- aggregating individual preference information with signals from the whole user population (as in collaborative filtering),
- the design of the user interface including the generation of explanations,
- the automated acquisition of the product model,
- compiling the product model to enable fast query answering and interactive search, as well as
- trust-related issues.
The aim of the workshop is to bring together researchers and practitioners that are working on topics related to the use of Artificial Intelligence techniques in the areas of recommendation systems, electronic commerce, personalised web tools, and interactive configurator systems. We are specifically interested in contributions which aim to combine explicit preference and constraint models with machine learning techniques.