| N13-1015 | performance using a relatively simple | backoff smoothing | method . The intuition behind |
| N01-1023 | tag dictionary ( used within a | backoff smoothing | strategy ) for labels are not |
| D13-1143 | For simplicity , we use stupid | backoff smoothing | ( Brants et al. , 2007 ) . 3 |
| P00-1048 | the formalism of the simplified | backoff smoothing | , each probability whose ML estimate |
| P04-1008 | statistical language model using Katz | backoff smoothing | technique ( Katz , 1987 ) . This |
| D12-1032 | characters are estimated with | backoff smoothing | . 5 Inference The input to inference |
| J14-4002 | using Witten-Bell smoothing , and | backoff smoothing | is achieved using failure transitions |
| P05-2004 | a lowerorder model , a form of | backoff smoothing | for dealing with data sparsity |
| P09-2022 | . Good-Turing discounting and | backoff smoothing | are also applied . Here , it |
| K15-1015 | features , giving us both a form of | backoff smoothing | and twenty times faster training |
| K15-1015 | faster training and a form of | backoff smoothing | . The resulting parser is over |
| D08-1113 | different lengths , a form of | backoff smoothing | ( Wu and Khudanpur , 2000 ) . |
| P10-1157 | using Witten-Bell interpolated | backoff smoothing | ( Bilmes and Kirchhoff , 2003 |
| N06-1036 | embedded EM training Incorporating | backoff smoothing | procedures into Bayesian networks |
| J05-4005 | lists using MLE , together with a | backoff smoothing | schema , as described in Section |
| P08-1085 | In all experiments , we use the | backoff smoothing | method of ( Thede and Harper |
| P06-1125 | . In our implementation , Katz | Backoff smoothing | technique ( Katz , 1987 ) is |
| P08-1058 | longer n-gram in such cases . | Backoff smoothing | algorithms typically request |
| P04-1021 | transliteration units ; 2 ) The | backoff smoothing | of n-gram TM is more effective |
| E97-1056 | analysed the relationship between | Backoff smoothing | and Memory-Based Learning and |
