| D15-1210 | measure the disagreement between | asymmetric alignments | . The loss functions can not |
| D15-1210 | Note that the union of the two | asymmetric alignments | is almost the same with the goldstandard |
| D15-1210 | encourages the two models to agree on | asymmetric alignments | . While this strategy significantly |
| D15-1210 | functions can be defined both between | asymmetric alignments | and between alignments and other |
| D15-1210 | can not only be defined between | asymmetric alignments | but also between alignments and |
| P11-2044 | 2010 ) . Although the absence of | asymmetric alignment | features affects performance |
| W05-0812 | a high-precision , low-recall | asymmetric alignment | , but symmetrization by intersection |
| N06-1014 | spaces . We first replace the | asymmetric alignments | a with a set of indicator variables |
| W15-1824 | ( Koehn et al. , 2007 ) . The | asymmetric alignments | are symmetrized with the intersection |
| D08-1084 | F1 . On the RTE2 test set , the | asymmetric alignment | from H to P scored 68 % in F1 |
| W05-0812 | -RSB- . 2.3 Does NULL Matter in | Asymmetric Alignment | ? Och and Ney ( 2000a ) introduce |
| Q13-1024 | one target word , and defines an | asymmetric alignment | for the sentence pair as 1 = |
| W05-0812 | NULL alignment is unnecessary for | asymmetric alignment | models when we symmetrize using |
| D15-1210 | alignment methods . 2 Background 2.1 | Asymmetric Alignment | Models Given a source-language |
| D08-1084 | GROW - DIAG-FINAL-AND -- and with | asymmetric alignments | in both directions . While all |
