P00-1034 simplified version of the well-known back-off smoothing method is used . To mitigate
P00-1034 simplified version of the well-known back-off smoothing method is used . To mitigate
J14-4003 Factored Language Models with back-off smoothing . Section 5 presents two methods
E97-1032 . Employing a small amount of back-off smoothing also for the known words is useful
J15-2001 suffer from data sparsity and the back-off smoothing may fall back to very short contexts
D13-1024 sequence . Models based on so-called back-off smoothing have shown good predictive power
C00-1070 lexicalized models use a simplified back-off smoothing technique to overcome data Sl
J04-2004 deal with unseen n-grams , the back-off smoothing technique from the CMU Statistical
C96-2151 idea is usually referred to as back-off smoothing , see ( Katz 1987 ) . These techniques
J14-4003 using Witten -- Bell interpolated back-off smoothing , according to the back-off graphs
J02-3004 no increase in performance for back-off smoothing ( see Tables 7 and 8 ) . These
C00-1070 the formalism of our simplified back-off smoothing , each of probabilities whose
J02-3004 necessary for class-based and back-off smoothing , we maintain the train/test
C00-1070 using the perplexity measure , a back-off smoothing ( Katz , 1987 ) is said to perform
J02-3004 1999 ) use perplexity to compare back-off smoothing against distance-weighted averaging
N13-1117 processing due to their relationship to back-off smoothing . Denoting a context of atoms
C94-1023 respectively . After applying back-off smoothing ( Katz 1987 ) and robust learning
D15-1231 This can be viewed as a kind of back-off smoothing ( Katz , 1987 ) , where the Observed
J02-3004 organization of conceptual information . Back-off smoothing , however , incorporates no notion
C96-2151 ' k - ~ , Tk-1 ) . 3.1 N-gram Back-off Smoothing We will first consider estimating
hide detail