probability model 0.0029846300000000003
parsing model 0.002794525
probabilistic model 0.002668623
discriminative model 0.002611529
same model 0.0025862750000000003
model parameter 0.0024922250000000003
classification model 0.0024569
generative model 0.002421508
model ssn 0.0023237
minimization model 0.002320504
bilistic model 0.002319621
tic model 0.002318206
model parame 0.00229945
ability model 0.00229945
probability feature 0.00214395
model 0.00205782
different feature 0.00185334
feature vector 0.001756563
training data 0.0016969910000000001
parsing models 0.0016708880000000001
kernel feature 0.001617743
other models 0.001586934
feature extraction 0.001572474
feature space 0.001567858
probabilistic models 0.001544986
parsing data 0.001539079
loss function 0.0015178029999999999
training set 0.0015014310000000001
feature spaces 0.001498283
loss learning 0.001484548
discriminative probability 0.001480519
feature extractor 0.001463603
feature weighting 0.001459429
different loss 0.0014038929999999998
probability candidate 0.00139926
kernel features 0.001390143
kernel models 0.0013347860000000001
classification models 0.001333263
classifier training 0.001325684
baseline models 0.001323459
linguistic features 0.001314005
function pos 0.001301945
learning algorithm 0.00129273
generative probability 0.001290498
loss estimation 0.001286099
decoding method 0.0012748360000000001
discriminative learning 0.001270564
probability estimates 0.001256783
parsing task 0.00125386
joint probability 0.001245035
loss candidate 0.001240143
learning task 0.00123401
previous models 0.001230167
feature 0.00121714
pcfg models 0.00121434
available training 0.001213962
ing set 0.0012127589999999999
machine translation 0.001207772
discriminative distribution 0.0012039820000000001
tree kernel 0.001199231
total probability 0.001193822
predefined probability 0.001192889
probability esti 0.001190493
probability estima 0.001181715
struct training 0.001180097
erative models 0.001178198
parsing problem 0.001177179
mum probability 0.001172719
inative probability 0.001172483
loss kernel 0.001168296
same method 0.001162892
other words 0.001157691
training criteria 0.001157405
loss approximation 0.001150709
classifier learning 0.001147922
first set 0.00113853
linear function 0.001134933
loss functions 0.001129566
parsing experiments 0.001129535
original data 0.001117294
word error 0.00110512
other accuracy 0.001075113
arbitrary loss 0.001073854
expected loss 0.001070587
loss kernels 0.001058283
word sequences 0.001055518
dard tree 0.001053554
tree ker 0.001052779
pected loss 0.0010476259999999999
the loss 0.001036253
learning algorithms 0.001035896
loss func 0.001035234
classification method 0.0010335169999999999
loss minimization 0.0010303769999999999
ing accuracy 0.001028307
loss logit 0.001021503
network learning 0.00101986
different approaches 0.0010154209999999999
journal parsing 0.001014755
wsj parsing 0.001010244
