Describes a system called LIEP that learns patterns for extracting information from raw text such as newswire articles, given examples of what to extract by end users