CS&E logo University of Washington Department of Computer Science & Engineering
 ARNAULD: Preference Elicitation For Interface Optimization
  CSE Home   AI Home  About CSE    Search    Contact Info 

Project faculty
 Daniel Weld
Project students
 Krzysztof Gajos

Document ArchiveCSE only
CVS HistoryCSE only
Java Docs

Last updated: April 3, 2005

Arnauld ARNAULD is named after Antoine Arnauld (1612-1694), a French philosopher who first applied the principle of Maximum Expeted Utility.


Recent years have revealed a trend towards increasing use of optimization as a method for automatically designing aspects of an interface's interaction with the user. In most cases, this optimization may be thought of as decision-theoretic -- the objective is to minimize the expected cost of a user's interactions or (equivalently) to maximize the user's expected utility. While decision-theoretic optimization provides a powerful, flexible, and principled approach for these systems, the quality of the resulting solution is completely dependent on the accuracy of the underlying utility or cost function. Unfortunately, determining the correct utility function is a complex, time-consuming, and error-prone task. While domainspecific learning techniques have been used occasionally, most practitioners parameterize the utility function and then engage in a laborious and unreliable process of hand-tuning.

Our work on Arnauld focuses on developing the interaction techniques and algorithms necessary to easily and robustly elicit the correct sets of parameters from the original developers or from the end users.

More Information and Publications

This extended abstract accompanying a UIST poster introduces the idea of automatically inferring the parameters of the objective function governing the behavior of SUPPLE from the user's responses to concrete interfaces. This abstract is subsumed by the UIST'05 paper.
Krzysztof Gajos, Raphael Hoffmann and Daniel S. Weld. Improving User Interface Personalization. In Supplementary Proceedings of UIST'04. Santa Fe, New Mexico, 2004
This full UIST paper presents the current state of Arnauld, a system that interactively elicits user preferences for the purpose of automatically learning parameters of optimization-based systems. We have used it to automatically generate the parameters of SUPPLE's cost functions.
Krzysztof Gajos and Daniel S. Weld. Preference Elicitation for Interface Optimization. In Proceedings of UIST'05, Seattle, WA, USA, 2005.
This ASSETS abstract outlines a new project that builds upon Supple and Arnauld. We are creating a system that allows non-experts to easily reparametrize the UI generator for the needs of individual users (potentially with physical or visual impairments), thus making automatic UI generation a scalable solution for providing people with interfaces custom tailored to their individual interaction needs and abilities.
Krzysztof Z. Gajos, Jing Jing Long and Daniel S. Weld Automatically Generating Custom User Interfaces for Users With Physical Disabilities In Proceedings of ASSETS'06, Portland, OR, 2006.

CSE logo Department of Computer Science & Engineering
University of Washington
Box 352350
Seattle, WA  98195-2350
(206) 543-1695 voice, (206) 543-2969 FAX
[comments to Krzysztof Gajos]