InfoRadar^TM was designed by the Programming Systems Research Group of the Laboratory for Computer Science at the Massachusetts Institute of Technology, and now is in the process of being updated by the Advanced Data Management Research Group of the PRECISE Project at the University of Puerto Rico at Mayagüez.

The central hypothesis of this work is that retrieval effectiveness of multilingual documents can be improved by simultaneously providing the search engine human-translated multilingual queries identified with their source languages. InfoRadar^ML enhances InfoRadar by adding support for multilingual queries and document collections.



Project Leader:
   Dr. Bienvenido Velez-Rivera

Students:
   Jairo E. Valiente-Fernandez (graduate)
   Lissete Toledo (undergraduate)



Description:

InfoRadar, an information retrieval system developed at the MIT Laboratory for Computer Science , supports a novel user interaction model based on Visual Query Hierarchies . Visual query hierarchies help users quickly focus on their particular information needs. Formulating precise and effective queries in information retrieval systems has always been a difficult task, even for experienced users. Several factors contribute to this problem. The task of formulating an effective query requires the user to predict which terms appear in documents relevant to the information need. This often requires extensive knowledge about the document collection that it difficult to obtain in large document corpora. In addition, users want to avoid retrieving irrelevant documents due to a query that is under-specified or contains ambiguous terms. As a result, users of an information retrieval system with a large corpus are often faced with the task of manually sifting through very large and often inappropriate result sets.

This research has proposed a possible solution to the critical issue of retrieving and analyzing of multilingual information resources on the University of Puerto Rico website system, using Inforadar^ML. The proposed prototype takes into account the requirements and constraints of each language supported(English and Spanish).

InfoRadar^ML enhances InfoRadar, our previous research prototype, by adding support for multi-lingual queries and document collections. The main purpose of the new enhancements is to allow users to formulate and process queries containing terms in multiple languages. These queries are sent to the Inforadar^ML search engine where they are matched against a collection of documents, also in multiple languages. The search engine then returns an integrated result set containing all documents in the selected languages. In order to fulfill this goal, we have focused our client-side implementation effort along two lines: providing users with ways to type special characters into their queries, and allowing users to submit multi-lingual queries to the search engine. On the server side, Inforadar^ML has been enhanced with a multi-lingual indexing module capable of automatically tagging documents with their source language and, based on this, conduct feature extraction using language-specific algorithms. The search engine proper has also been enhanced with multi-lingual query processing capabilities.


Experimental Plan:

• Implement Inforadar site indexing ALL website data at UPRM
• Develop a multilingual support
• Make Inforadar the official search engine for the UPRM web site
• Conduct usability study
• Analyze real user feedback
• Incorporate feedback into an improved design

[ Next: How Inforadar works ]