Stanford Digital Library Project

SIDL-WP-1995-0020

LIRA II: A Plan for the Next Phase of Work on the LIRA System

Marko Balabanovic

marko@cs.stanford.edu

Abstract: An increasingly important problem is coping with the volume of information available on-line. This is especially true of the world-wide web, which now encompasses at least 5 million pages (as of July 15th the Lycos system (http://www.lycos.com) had indexed 5,077,834 unique URLs). Systems have been designed to alleviate these problems by filtering or routing streams of incoming information (such as email or news articles), and by creating indexes which allow users to search for specific information. In contrast our approach has been to create a system (called LIRA---Learning Information Retrieval Agent) which assists with browsing by delivering interesting world-wide web pages. In this report I will discuss plans for the new LIRA II system which will attempt to overcome the problems discovered with the original LIRA system. At this stage I am mainly proposing a platform which will enable many different experiments, allowing us to find out which of the possible design variants will work best. At the same time the platform is intended to be a working system (the experiments will involve real users), so I will suggest a particular design for the required infrastructure.

Note: Papers in this series are in development and are not in a final form for publication or general dissemination. They are subject to change. Please do not quote or further distribute them without explicit permission from the authors.

This paper was created on: 08/10/95 and last revised on:10/19/1995

Author's Comments: This report explains my current high-level design for a new version of the adaptive information retrieval system for the world-wide web.

Status: PRIVATE

Click here to see the full text of SIDL-WP-1995-0020 (PS)