language model 0.0029938200000000003
training data 0.0026467419999999997
test data 0.002396055
exact model 0.002341743
model constant 0.002321352
window model 0.002272171
model order 0.002266899
fixed model 0.002226826
reservoir data 0.002214312
guage model 0.002148518
model affect 0.00213854
streaming data 0.0021196739999999998
twitter data 0.0020391429999999998
data type 0.002024988
unseen data 0.002021695
available data 0.002012104
data structures 0.0019888099999999997
language models 0.00197217
textual data 0.001961498
data does 0.0019479019999999998
model 0.00190573
sampling function 0.0015596429999999999
large language 0.00150953
sampling probability 0.00148522
sampling algorithm 0.001476784
new training 0.0014192839999999998
statistical language 0.0014020670000000001
modeling language 0.001395972
language modelling 0.001350763
reservoir sampling 0.00133838
current training 0.001333821
same test 0.00133029
language changes 0.001315446
uniform sampling 0.0012654839999999999
test set 0.0012431830000000001
sampling rate 0.001231968
same results 0.001227543
window sampling 0.001212489
exponential sampling 0.0012086039999999998
sampling methods 0.001205876
topic models 0.001198523
alternative sampling 0.0011828469999999999
following models 0.001172918
different sample 0.0011606820000000001
rolling training 0.001156945
incredible feature 0.001139434
constant function 0.001129217
building models 0.0011241839999999999
sampling case 0.001116124
trigger models 0.001110574
probability distribution 0.001105443
tial sampling 0.001099466
sampling strategies 0.001095191
language 0.00108809
arbitrary sampling 0.001085725
same number 0.001084458
current test 0.001083134
precise sampling 0.001073833
sampling strat 0.001072146
algorithm parameters 0.001048638
such biases 0.001015675
decay function 9.88598E-4
constant space 9.742290000000001E-4
randomised algorithm 9.69586E-4
same amount 9.67115E-4
test point 9.66998E-4
stream size 9.639410000000001E-4
standard reservoir 9.50834E-4
reservoir sample 9.48207E-4
different biases 9.47901E-4
different age 9.46121E-4
large stream 9.437180000000001E-4
window results 9.37769E-4
efficient algorithm 9.35374E-4
same dis 9.34987E-4
different problems 9.34636E-4
reservoir size 9.33995E-4
word corpus 9.31319E-4
training 9.24762E-4
previous stream 9.16603E-4
uniform reservoir 9.117679999999999E-4
feature 9.11709E-4
perplexity results 9.046709999999999E-4
translation system 8.98728E-4
sample size 8.97538E-4
see algorithm 8.94344E-4
machine translation 8.92628E-4
same rolling 8.883980000000001E-4
other experiments 8.86937E-4
uniform distribution 8.857069999999999E-4
random sample 8.84184E-4
models 8.8408E-4
fixed space 8.79703E-4
equal probability 8.75713E-4
uniform sample 8.753109999999999E-4
new words 8.6702E-4
large values 8.59553E-4
exponential reservoir 8.548879999999999E-4
experimental results 8.54539E-4
dynamic stream 8.503580000000001E-4
