As the Web provides rich data embedded in the immense contents inside pages, we witness many ad-hoc efforts for exploiting fine granularity information across Web text, such as Web information extract