Transforming Examples into Patterns for Information Extraction

Information Extract ion (IE) systems today are commonly based on pat tern matching. The pat terns are regular expressions stored in a customizable knowledge base. Adapting an IE system to a new subject domain entails the construction of a new pat tern base a t ime-consuming and expensive task. We describe a s trategy for building pat terns from examples. To adapt the IE system to a new domain quickly, the user chooses a set of examples in a training text, and for each example gives the logical form entries which the example induces. The system transforms these examples into pat terns and then applies meta-rules to generalize these patterns.