Automated metadata and instance extraction from news Web sites | IEEE Conference Publication | IEEE Xplore