InCommonSense - presenting search-engine results with retrieved human authored summaries

Einat Amitay,  einat@mri.mq.edu.au
MRI, School of MPCE, Macquarie University

Tuesday 23 March, at 11am

Abstract

It is common knowledge that many users are overwhelmed by the amount of information returned by web search engines. Many of them complain that there is a need to 'visit' each result in order to find out whether or not it is relevant. This process is tiring and time consuming, which is probably why some people find using search engines a difficult task. Although commercial search engines attempt to give the title and the first lines of the document for each search result, there is very often insufficient information to assess the relevancy of the result without going to the document itself and reading it.

This talk would suggests a new way for automatically retrieving and reusing human generated annotations on the web, incorporating these annotations into a search engine for creating a hybrid directory-search-engine which allows for both automatic retrieval and on-the-fly human authored summaries.

A model will be presented for reusing human annotations on the web for describing web search engine results. The need for coherent summaries for web search results will be discussed and some examples for such summaries will be presented.

A large scale experiment which tries to study people's preferences regarding these kind of summaries will be reported and some of the results will be shown.

The system implementing this model can be viewed under:
http://www.mri.mq.edu.au/~einat/incommonsense/

 

Short resume

Einat Amitay is a PhD student at the Microsoft Research Institute, Macquarie university. Her work is sponsored by the Microsoft Research Institute and CSIRO Mathematical and Information Sciences, and supervised by Cecile Paris from CSIRO, and Jon Oberlander from the division of Informatics, Edinburgh, Scotland.

Einat holds a BA in French and English linguistics from the Haifa university, Israel; and an MSc from The Centre for Cognitive Science, Edinburgh, Scotland. Her main research interests are defining language-use conventions in web documents and using these to identify descriptions of other web documents.

 

Back to HAIL Home Page