W06-3108 defines an event for the maximum entropy training . An exception are the oneto-many
W06-1646 prohibitively expensive for maximum entropy training . Analysis of the models learned
W11-3214 Pairs The features for maximum entropy training are extracted from aligned names
P08-1011 Goodman , 1998 ) . The Maximum Entropy training toolkit from ( Zhang , 2006 )
W06-1646 feature selection and the maximum entropy training procedure . 8 Acknowledgements
E03-1007 baseline system and from the maximum entropy training on the transformed corpus . For
W11-1007 weights obtained through the maximum entropy training on the parallel data . Finally
W10-4133 some features for the maximum entropy training . However , it effectively improves
N04-1039 2001 . Classes for fast maximum entropy training . In ICASSP 2001 . Joshua Goodman
W11-2708 ) ( 4 ) Given that the maximum entropy training procedure attempts to minimize
P02-1038 five-gram GIS algorithm for maximum entropy training of alignment templates . language
W06-3108 formance . Here , we let the maximum entropy training decide which features are important
W07-1516 Our implementation of maximum entropy training employs a convex optimization
D12-1095 subtree ranker method using Maximum Entropy training ( 's ubtree ranking by Max -
W09-2902 Again , supervised denotes Maximum Entropy training and Unsupervised is our unsupervised
W10-1410 Chrupała et al. ( 2008 ) use Maximum Entropy training to learn PM and PL , here we
E09-1033 more difficult . Their Maximum Entropy training is more appropriate for their
P07-1091 alignment tool , and a Maximum Entropy training tool . We use the Stanford parser
P02-1038 translations used for the maximum entropy training . • WER ( word error rate
P02-1038 problem , we define for maximum entropy training each sentence as reference translation
hide detail