Showing papers by "Eugene Agichtein published in 2001"

Learning search engine specific query transformations for question answering

Abstract: We introduce a method for learning query transformations that improves the ability to retrieve answers to questions from an information retrieval system. During the training stage the method involves automatically learning phrase features for classifying questions into different types, automatically generating candidate query transformations from a training set of question/answer pairs, and automatically evaluating the candidate transforms on target information retrieval systems such as real-world general purpose search engines. At run time, questions are transformed into a set of queries, and re-ranking is performed on the documents retrieved. We present a prototype search engine, Tritus, that applies the method to web search engines. Blind evaluation on a set of real queries from a web search engine log shows that the method significantly outperforms the underlying web search engines as well as a commercial search engine specializing in question answering.

Snowball: a prototype system for extracting relations from large text collections

Abstract: Text documents often hide valuable structured data. For example, a collection of newspaper articles might contain information on the location of the headquarters of a number of organizations. If we need to nd the location of the headquarters of, say, Microsoft, we could try and use traditional information-retrieval techniques for nding documents that contain the answer to our query. Alternatively, we could answer such a query more precisely if we somehow had available a table listing all the organization-location pairs that are mentioned in our document collection. One could view the extraction process as automatically building a materialized view over the unstructured text data. In this demo we present an interactive prototype of our Snowball system for extracting relations from collections of plain-text documents with minimal human participation. Our method builds on the DIPRE idea introduced by Brin [3]. Our system and techniques were presented in detail in [2] and [1].