Workshop on Web Search Result Summarization and Presentation
Co-located with the 18th World Wide Web Conference
April 20th, 2009, Madrid, Spain
Providing a satisfying web search experience can be a challenging task for a search engine. Numerous disciplines -- search, summarization, user interface design, usability, metrics, machine learning and modeling -- all have to come together in order to deliver the final experience. Effective summarization is part of the challenge. In this workshop we will focus on various aspects of web summarization, presentation, and user satisfaction metrics and models. The kinds of questions and issues we would like to address are:
- What makes an effective web search result summary? What is a summary for?
- Technological challenges and opportunities for innovation in how summaries are generated
- What should be optimized during summarization?
- Defining and measuring effectiveness of summarization
- What real-time measurements (e.g. click logs) and offline (e.g. human rater judgments) can be used as surrogates for user satisfaction models?
- How does one model the differences in human scanning and reading behaviors?
- How can eye tracking technology be utilized to help understand the balance between scanning and reading behaviors?
- What useful qualitative insights can we gauge from usability and field studies with respect to how users utilize summaries in their search for information?
- What are good scalable metrics for summarization?
- Can one learn layout and presentation using appropriate machine learning techniques and targets?
- How do we optimize SERP UI to support user workflow?
- What's good and not good about presenting ranked results linearly?
- Do new presentation strategies overcome the limitations?
- Future trends and directions in search results presentation
Main topics of interest include but are not limited to:
- Web summarization and related natural language processing
- Information presentation, exploration, and design
- Usability and eye tracking studies of web search results presentation
- Machine learning for summarization and presentation
- User models: learning from clicks and human rater judgments
- Metrics for individual pieces and the final experience
Submission deadline: 20th February 2009