P02-1035 provide partially labeled data for discriminative estimation of a probability model on linguistically
D10-1069 a difference of generative vs. discriminative estimation . We suspect that the latent
W03-1013 These techniques are necessary for discriminative estimation techniques applied to wide-coverage
W12-2706 3.3 ) . The second issue is that discriminative estimation of LMs is computationally more
P04-1014 Riezler et al. ( 2002 ) in using a discriminative estimation method by maximising the conditional
P02-1035 exploitation of the WSJ treebank for discriminative estimation of an exponential model on LFG
P02-1035 Lexical-Functional Grammar and Discriminative Estimation Techniques </title> Stefan Riezler
J07-4004 the parse space available for discriminative estimation , which is possible on the cluster
P04-1007 kind of discriminative training . Discriminative estimation of language models has also been
N06-1019 dependency model , which uses a discriminative estimation method by maximising the conditional
W04-3223 ( 2004 ) ) . For the setup of discriminative estimation from partially labeled data ,
N10-1071 function ` llm that was used for discriminative estimation from sets of partially labeled
P02-1035 WSJ training sections to guide discriminative estimation . That is , discriminative criteria
D11-1148 were inspired by earlier work in discriminative estimation techniques , such as Johnson
P10-1130 their idea has been extended to discriminative estimation ( Rie - zler et al. , 2002 )
P04-1007 vocabulary ASR has benefitted from discriminative estimation of Hidden Markov Model ( HMM
P02-1035 the corresponding strings in a discriminative estimation setting . P ( &#955; ) = &#8722;
P09-1042 p&#952; ( z | x ) ) -RSB- . For discriminative estimation ( Ganchev et al. , 2008 ) , we
J07-4004 is twofold . First , it makes discriminative estimation feasible by limiting the number
J07-4004 Riezler et al. ( 2002 ) in using a discriminative estimation method by maximizing the conditional
hide detail