Title:   Indexing Shared Content in Information Retrieval Systems 
Speaker: Marcus Fontoura (Yahoo! Research)


Abstract:

In many corpora, documents share some content in full or in
part. However, most information retrieval systems process each document
separately, causing shared content to be indexed multiple times. In this
paper, we describe a new document representation model where related
documents are organized as a tree, which facilitates single indexing of
shared content.  We show how this representation can be encoded in  an
inverted index and we devise algorithms for evaluating free-text queries
based on this encoding. We also show how our representation applies to web,
email, and newsgroup search.  Finally, we present experimental results that
indicate that our method can provide a significant reduction in the size of
an inverted index as well as in the time to build and query it. 


Short-Bio:

Marcus Fontoura joined Yahoo! Research in November 2005. Prior to
this, he was a Research Staff Member at the IBM Almaden Research Center in
the Computer Science Department. He also had research posts at the Computer
Systems Group, University of Waterloo and at Princeton University. His
primary interests are web search, data management, algorithms and data
structures for large data sets. He obtained his PhD in Computer Science from
PUC-Rio, Brazil in July 1999.