model size 0.00196351
markov model 0.001942518
model selection 0.001899848
model minimization 0.001879498
approximate model 0.001871416
model simultane 0.001865038
model 0.00161891
word sequence 0.001561361
word type 0.001520832
data set 0.001266646
test data 0.001247868
words figure 0.0012429329999999999
english pos 0.001226268
other techniques 0.001193622
pos tagging 0.001192315
function value 0.001188499
test set 0.001178902
unique tag 0.001152971
other tasks 0.0011525160000000001
objective function 0.00114216
other prob 0.001130009
search algorithm 0.00108412
other problems 0.001078936
common method 0.001074593
tag bigram 0.001066811
corpus probability 0.001062143
tag vocabulary 0.0010513739999999999
tagging accuracy 0.001029944
data sets 0.001019564
jective function 0.001012391
pos tags 0.001011525
generative models 0.001003863
recent work 0.001000713
tagging problem 9.94512E-4
bayesian approach 9.90423E-4
posterior parameter 9.89912E-4
parameter optimization 9.728240000000001E-4
unsupervised pos 9.68121E-4
tagging task 9.46887E-4
italian data 9.414429999999999E-4
available data 9.35613E-4
unlabeled data 9.32263E-4
italian corpus 9.156800000000001E-4
possible tags 9.014870000000001E-4
tive models 8.98448E-4
large improvement 8.97062E-4
italian pos 8.967860000000001E-4
parameter vector 8.923850000000001E-4
high pos 8.90701E-4
latter approach 8.810840000000001E-4
general search 8.737479999999999E-4
viterbi tagging 8.71443E-4
tagging experiments 8.60658E-4
linear program 8.54549E-4
bigram hmm 8.52479E-4
system accuracy 8.52467E-4
zero parameters 8.47888E-4
same experiment 8.46573E-4
average accuracy 8.38313E-4
words 8.35362E-4
size term 8.28897E-4
tagging accuracies 8.27297E-4
parameter estimate 8.21591E-4
integer linear 8.195769999999999E-4
tagging accu 8.152159999999999E-4
parameter estimates 8.12469E-4
constant term 8.06464E-4
gibbs sampling 8.02692E-4
ous work 7.998860000000001E-4
efficient search 7.97154E-4
equality constraints 7.790119999999999E-4
hidden markov 7.62816E-4
search procedure 7.61774E-4
ging accuracy 7.61227E-4
bayesian methods 7.547999999999999E-4
function 7.51359E-4
method 7.3214E-4
optimization prob 7.312480000000001E-4
term logz 7.30643E-4
erative search 7.29422E-4
bayesian inference 7.25848E-4
likelihood maximization 7.24688E-4
optimization algorithms 7.19226E-4
likelihood estimation 7.13478E-4
maximum likelihood 7.054590000000001E-4
generic way 7.04135E-4
science division 7.00481E-4
constant terms 6.99338E-4
polynomial time 6.892420000000001E-4
problem formulation 6.86669E-4
system zero 6.837760000000001E-4
results 6.80046E-4
probability distri 6.78026E-4
overfitting problem 6.76226E-4
optimization tool 6.74422E-4
types maximum 6.69655E-4
channel probability 6.67192E-4
norm yields 6.64951E-4
admiralty way 6.59788E-4
penn treebank 6.59205E-4
