C04-1109 preprocessed by the Proteus NE tagger and Charniak sentence parser . Then the GLARF processor produced
P05-1052 input text is preprocessed by the Charniak sentence parser ( including tokenization and
hide detail