International Business Machines Corporation

Rani Nelken

Aya Soffer

Einat Amitay

David C Smith

Niblack Wayne

A method for extracting information from a corpus of data includes specifying a topic and a query term associated with the topic, and defining adjunct terms which may occur in the corpus in a context of the query term, the adjunct terms comprising one or more off-topic terms. Occurrences of the query term are found in the corpus, the occurrences including at least one occurrence of the query term together with at least one of the off-topic terms in the context of the query term. The at least one occurrence of the query term is classified as non-relevant to the topic responsively to the occurrence of the at least one of the off-topic terms in the context of the query term.

Growing community of inventors

Cambridge, MA, United States of America

Rani Nelken