W04-3216 to add " fake counts " during parameter re-estimation , according to the prior . The
P11-2124 steps is followed by an EM-based parameter re-estimation . This process allows learning
A00-2014 Baum-Welch Maximum Likelihood parameter re-estimation on diagonal covariance Gaussian
J05-4004 The final inference problem is parameter re-estimation . In the case of standard HMMs
P98-1035 useful for justifying the model parameter re-estimation . The two estimates ( 8 ) and
H93-1020 are reported on the benefits of parameter re-estimation . For example , while many researchers
J12-3007 algorithm for a second stage of parameter re-estimation for WORD-PREDICTOR and SEMANTIZER
D15-1119 of word orders and an efficient parameter re-estimation algorithm is devised . It has
P11-1021 algorithm for a second stage of parameter re-estimation for WORD - PREDICTOR and SEMANTIZER
P99-1022 reduction using document-specific parameter re-estimation , and no significant word error
P98-1035 techniques similar to those used in HMM parameter re-estimation can not be used with our model
J05-4004 technical requirement involving parameter re-estimation , which essentially says that
J12-3007 31 ) . We use a second stage of parameter re-estimation for p ( wk +1 | wkk − n
P07-1036 from the supervised model . The parameter re-estimation in line 9 , uses a similar intuition
W01-0505 ) . In other works , iterative parameter re-estimation ` Suppose that Uk and vk are
P98-1035 perplexity of our model . 3.4 Parameter Re-estimation The major problem we face when
hide detail