Classical Information Extraction (IE) systems fill slots in domain-specific frames.
Classical IE systems fill slots in domain-specific frames such as the time and location slots in seminar announcements (Freitag, 2000) or the terrorist organization slot in news stories (Chieu et al., 2003).
Sequence extraction has two parts: identifying possible extractions (at, k, s) from text, and then classifying those extractions as either correct or incorrect.
This section reports on two experiments.
There has been extensive work in extracting lists or sets of entities from the Web.
We have demonstrated that an extractor leveraging sequence regularities can greatly outperform extractors without this knowledge.