segmentation word 0.00468194
word segmentation 0.00468194
chinese word 0.004330553
word sequence 0.004259219
same word 0.004183316
joint word 0.004163925
word probability 0.0040652959999999995
sentence word 0.004035756
input word 0.003943658
word seg 0.003942791
word lattice 0.003879182
word boundaries 0.003873562
word probabilities 0.00386987
word sequences 0.0037912189999999998
word segmentations 0.003788322
word segmenta 0.003784982
weighted word 0.003778833
word offsets 0.003778746
word spanning 0.0037779249999999997
our word 0.003776777
compacted word 0.003776777
pos tag 0.00269325
chinese pos 0.002533693
pos tags 0.002529827
pos tagging 0.002405567
pos tagger 0.002321255
chinese character 0.002161513
sequential pos 0.002002869
segmentation system 0.001843174
same feature 0.0018344060000000002
segmentation results 0.0017891909999999999
identical words 0.001713223
parsing system 0.001706354
character offsets 0.001609706
syntactic parsing 0.001594657
feature templates 0.001569861
segmentation evaluation 0.001548542
entropy model 0.001516091
inference parsing 0.0015154650000000001
standard data 0.0014788000000000002
segmentation sys 0.001463396
feature weights 0.001460959
treat segmentation 0.001455558
tag set 0.001445647
words 0.00143713
feature tem 0.001429818
parsing systems 0.00142349
chinese tree 0.001412029
learning approach 0.001370394
chinese sentence 0.001363189
joint learning 0.001346126
character 0.00133252
tactic parsing 0.001319955
decoding algorithm 0.0012803110000000001
training set 0.001273284
chinese treebank 0.001244331
previous work 0.0012245609999999999
chinese experiments 0.001213689
tree figure 0.0012037559999999998
data split 0.001195852
segmentation 0.00118038
similar decoding 0.001163818
high accuracy 0.001161388
inference approach 0.001158538
model 0.00115708
feature 0.00115265
chinese sen 0.001148639
search space 0.001141883
joint task 0.001136233
joint inference 0.0011342700000000002
perceptron algorithm 0.001133485
ing system 0.001112107
good search 0.0011103229999999999
chinese academy 0.0011041129999999999
decomposition method 0.001102859
entire training 0.001094552
different strengths 0.001094423
pipeline approach 0.00109281
subgradient method 0.0010890980000000001
research problem 0.001086966
line approach 0.001083157
ing performance 0.0010801629999999999
combined search 0.0010706700000000001
parse tree 0.001069861
tagging evaluation 0.001069029
pipeline system 0.001068971
sequence labeling 0.0010655740000000001
parser seg 0.001064611
syntactic structure 0.0010507939999999999
enough search 0.001049325
decoding algorithms 0.001045365
parsing 0.00104356
sequence brown 0.0010373700000000001
pipeline parser 0.001029557
validation approach 0.0010142419999999998
different viewpoints 0.00100489
framework figure 0.001002277
different perspectives 0.001000639
original decoding 9.95012E-4
mentation system 9.91331E-4
