The HAIRCUT System for Cross-Language Information Retrieval – James Mayfield (Johns Hopkins University Applied Physics Laboratory)
View Seminar Video
Cross-language information retrieval is the retrieval of documents in one language that are relevant to a query in another language. The Hopkins Automated Information Retriever for Combing Unstructured Text (HAIRCUT) system is an information retrieval engine that exhibits exceptional performance on the cross-language retrieval task. In this talk, I will first give an overview of information retrieval, and describe the HAIRCUT approach. Next, I will show how HAIRCUT uses combination of evidence to exploit the strengths of various types of indexing terms and similarity metrics. Finally, I will describe how HAIRCUT performs cross-language retrieval, and examine its strengths and weaknesses in this task.
James Mayfield is a Senior Computer Scientist at the Johns Hopkins University Applied Physics Laboratory. His doctoral work at the University of California at Berkeley was on plan recognition in dialogue systems. Since that time, he has worked in areas as diverse as hypertext systems, information extraction, and agent communication languages. Prior to joining JHU/APL, Dr. Mayfield was an associate professor of computer science at the University of Maryland Baltimore County (UMBC).