| D08-1041 | thousands of high precision bilingual | name translation | spelling variants . We applied |
| D08-1041 | of thousands of high precision | name translation | pairs . We further apply these |
| D08-1041 | Translation We applied the extracted | name translation | spelling variants to the machine |
| D08-1041 | can be trained from bilingual | name translation | pairs . We segment the source |
| D08-1041 | of thousands of high precision | name translation | spelling variants are extracted |
| D10-1042 | approach complements existing | name translation | work , by not requiring rare |
| D08-1041 | Ahmad Emami Imed Abstract Foreign | name translations | typically include multiple spelling |
| D08-1041 | selected threshold are considered as | name translations | . 2.1 Name Transliteration Cost |
| D08-1041 | Arabic-English and English - Arabic | name translation | variants from sentencealigned |
| C04-1127 | of an unknown word . In fact , | name translation | of the MT system we used for |
| D08-1041 | edit distance metric , target | name translations | with similar spellings are clustered |
| D08-1041 | transliteration costs are considered as | name translations | , and the target cluster contains |
| C04-1103 | extending our approach to handle such | name translation | . We also extending our method |
| D08-1041 | The accuracy of the extracted | name translation | spelling variants are judged |
| D08-1041 | experiment , manually correcting the | name translation | errors in the first 10 documents |
| A97-2018 | translation system to prevent | name translation | mistakes . Versions are available |
| D10-1042 | Conclusion This paper abstracted | name translation | problem as a matching problem |
| D08-1041 | Although the significance of correct | name translation | can not be fully represented |
| D10-1042 | the proposed method for mining | name translations | in Chinese and English languages |
| D08-1041 | the precision of the top 22K A-E | name translations | is 96.9 % . Among them 98.5 % |
