Special Issue of the
Artificial Intelligence Journal
Representing, Learning, and
Theoretical and Practical Challenges
Carmel Domshlak (Israel Institute
Eyke Hüllermeier (University of Marburg, Germany, firstname.lastname@example.org)
Souhila Kaci (IUT de Lens, France, email@example.com)
Henri Prade (University of Toulouse, France, firstname.lastname@example.org)
Background and Scope
The topic of preferences has recently attracted considerable attention in Artificial Intelligence (AI) and plays an increasingly important role in several AI-related research fields, including, e.g., agents, constraint satisfaction, decision theory, planning, machine learning, and argumentation. Representing and processing knowledge in terms of preferences appears to be especially appealing from an AI perspective, notably as it allows one to specify desires in a declarative way, to combine qualitative and quantitative modes of reasoning and to deal with inconsistencies and exceptions in a quite flexible manner.
Even though methods for dealing with preferences in a formal way have been developed in research areas such as operations research, game and decision theory, and social choice for quite a while, AI research has made novel and complementary contributions to preference handling during the last decade. Besides, having contemporary application fields such as electronic auctions, e-commerce, and recommender systems in mind, AI research has raised new problems and put emphasis on additional aspects such as combinatorial structure of the alternatives, algorithmic issues and complexity, and the absence of a decision analyst in the loop (which is typical for many applications).
Despite the existence of some established representation frameworks that allow for building and handling preference models in an effective and tractable way, research in the preference field has remained very active and still faces many challenges, such as comparing the expressivity and complexity of existing modeling languages, combining different representation modes and frameworks, supporting the knowledge acquisition task by discovering, learning, and adapting user preferences based on different types of feedback, handling preferences in multi-agent systems and group decision making, and developing efficient and theoretically sound algorithms for preference aggregation and revision, just to mention a few.
The aim of this special issue is to provide an up-to-date picture of the current trends in the AI research on handling user preferences. Especially welcome are contributions that bridge the gap between the rather particulate approaches existing so far, thereby helping to establish a coherent theoretical foundation of preferences in AI, as well as interdisciplinary work that combines or integrates approaches from other fields, such as operations research, databases, or game and decision theory.
All manuscripts must adhere to the submission guidelines of the AI journal (which can be found here) and should be submitted to the journal in the usual way (that is, not to the guest editors directly). To give notice of submitting to the special issue, please put "Special Issue on Preferences" in the subject heading of your email to email@example.com.
We kindly ask prospective authors to express, as far as possible,
their intention to submit a paper. To this end, please send an email
including a tentative title and a short abstract to all guest editors.
- Intention: December 15th, 2008
- Submission: February 15th, 2009
- Notification: June 15th, 2009
- Revised papers: July 30th, 2009