jz@cs.rmit.edu.au

Username: jz@cs.rmit.edu.au

Participant details

	Network Member	Justin Zobel
	Organisation/Institution	RMIT University
	Department/Centre	School of Computer Science & Information Technology
	Lab
	Website	http://goanna.cs.rmit.edu.au/~jz/
	Research area	Information Retrieval
	Research keywords

Research Interests

My research interests of relevance to this project are methodologies and algorithms for information retrieval. One of the key problems of information retrieval is accurate matching of user queries to relevant documents in the presence of confounding factors. These factors include scale, as data volumes may be in the terabytes; vocabulary, as author and user may use different terminology or even different languages; ambiguity, as queries have no textual context; and noise, as the documents may contain information that does not describe their content. My current research into solutions for these problems is based on using new statistical approaches to resolve ambiguity. For example, in cross-lingual retrieval, the documents in which the query words occur can be used to choose between alternative translations of the words. Matching of documents by topic can be used to determine whether alternative terminologies have been used for the same concept. Thus this work concerns text comprehension and adaptation to variability of input.