| D08-1050 | from the packed chart , using a | loglinear | model to score the derivations |
| C04-1006 | superior to the performance of the | loglinear | interpolation . In Table 4 , |
| C04-1006 | task , the improvement with the | loglinear | interpolation is statistically |
| D08-1076 | , 2008 ) . Each system uses a | loglinear | combination of 20 to 30 feature |
| C04-1006 | experiments , we investigated also the | loglinear | interpolation of the lexicon |
| C04-1006 | propose and justify linear and | loglinear | interpolation methods . Statistical |
| C04-1006 | significant at the 99 % level . For the | loglinear | interpo - lation , the target-to-source |
| C04-1006 | Canadian Hansards task . Here , the | loglinear | interpolation performs best . |
| C04-1006 | see that both the linear and the | loglinear | lexicon symmetrization methods |
| A00-2018 | This is useful if one is using a | loglinear | model for smoothing . That is |
| C04-1168 | in section 2 were used in the | loglinear | models . We did not perform parameter |
| C04-1006 | article Model 6 is introduced as the | loglinear | interpolation of the other models |
| C04-1180 | 2004b ) parser , which uses a | loglinear | model of normal-form derivations |
| C04-1006 | 0:5 for both the linear and the | loglinear | interpolation , i.e. both translation |
| C04-1006 | interpolation . Performing the | loglinear | interpolation , we observe a |
| C04-1006 | one direction is large . For the | loglinear | interpolation , the combined |
| C04-1006 | shown that both the linear and the | loglinear | interpolation of lexicon counts |
| D08-1091 | and Gao ( 2007 ) . Fitting the | loglinear | model involves the following |
| C04-1006 | polationresemblesmoreaunionofthetwolex - ica whereas the | loglinear | interpolation is more similar |
| C04-1032 | and Ney , 2003 ) . Model 6 is a | loglinear | combination of the IBM-4 , IBM-1 |
