| D11-1148 | specific purpose of augmenting the | Charniak parser | . However , much subsequent work |
| C04-1021 | this paper . We found that the | Charniak parser | , which was trained on the WSJ |
| C04-1021 | Parser Enhancements We used the | Charniak parser | ( Charniak , 2000 ) for the experiments |
| C04-1021 | for re-training . Whereas the | Charniak parser | was originally trained on close |
| D10-1031 | trees and the output from the | Charniak parser | , and there is not a big performance |
| D09-1161 | This is because that the enhanced | Charniak parser | provides more accurate model |
| D08-1093 | the more accurate , but slower | Charniak parser | ( Charniak and Johnson , 2005 |
| C04-1021 | great . Initially , plugging the | Charniak parser | into PRECISE yielded only 61.9 |
| D08-1050 | retraining the parser model . Since the | Charniak parser | does not use a lexicalized grammar |
| D11-1066 | stituency trees , we used the | Charniak parser | ( Char - niak , 2000 ) . We also |
| D09-1161 | Table 9 and Table 10 show that the | Charniak parser | enhanced by re-ranking and self-training |
| D09-1108 | source side parsed by a modified | Charniak parser | ( Charniak 2000 ) which can output |
| D09-1161 | confidence score and " C " means | Charniak parser | confidence score . 5.5 Comparison |
| D09-1161 | score output from the enhanced | Charniak parser | . Table 9 and Table 10 show that |
| D09-1161 | Berkeley parser and the enhanced | Charniak parser | by using the new model confidence |
| D08-1032 | well as for the query using the | Charniak parser | and measured the similarity between |
| D09-1108 | can output a packed forest . The | Charniak Parser | is trained on CTB5 , tuned on |
| D08-1093 | available answer is to take the | Charniak parser | performance on WSJ section 24 |
| D08-1093 | ) that was used for predicting | Charniak parser | behavior on the Brown corpus |
| D11-1096 | constituency trees , we used the | Charniak parser | ( Charniak , 2000 ) whereas we |
