Statistical machine translation processing

Abstract

A method of statistical machine translation (SMT) is provided. The method comprises generating reordering knowledge based on the syntax of a source language (SL) and a number of alignment matrices that map sample SL sentences with sample target language (TL) sentences. The method further comprises receiving a SL word string and parsing the SL word string into a parse tree that represents the syntactic properties of the SL word string. The nodes on the parse tree are reordered based on the generated reordering knowledge in order to provide reordered word strings. The method further comprises translating a number of reordered word strings to create a number of TL word strings, and identifying a statistically preferred TL word string as a preferred translation of the SL word string.

1 . A method comprising:
 generating, by a computer, a plurality of word strings corresponding to a syntactic tree structure that represents a source word string in a natural source language; and   selecting a preferred translation from translations of the plurality of word strings, the translations resulting from translating the plurality of word strings into a natural target language, the translating based on a plurality of alignment matrices that map sample source sentences in the natural source language to sample target sentences in the natural target language.   
     
     
         2 . The method of  claim 1  further comprising reordering, according to the alignment matrices, terms of the generated plurality of word strings with regard to a formal syntax. 
     
     
         3 . The method of  claim 2  wherein the translating is performed after the reordering. 
     
     
         4 . The method of  claim 2  wherein the reordering is further based on reordering probabilities calculated for each of the reordered plurality of word strings. 
     
     
         5 . The method of  claim 1  wherein the selecting is based on scores assigned to the translations of the plurality of word strings. 
     
     
         6 . The method of  claim 5  further comprising generating the assigned scores based on a log-linear paradigm. 
     
     
         7 . The method of  claim 1  further comprising building translation knowledge prior to the generating and selecting. 
     
     
         8 . At least one memory device storing instructions that, when executed by a computer, cause computer to perform a method of statistical machine translation (SMT), the method comprising:
 generating, by a computer, a plurality of word strings corresponding to a syntactic tree structure that represents a source word string in a natural source language; and   selecting a preferred translation from translations of the plurality of word strings, the translations resulting from translating the plurality of word strings into a natural target language, the translating based on a plurality of alignment matrices that map sample source sentences in the natural source language to sample target sentences in the natural target language.   
     
     
         9 . The at least one memory device of  claim 8 , the method further comprising reordering, according to the alignment matrices, terms of the generated plurality of word strings with regard to a formal syntax. 
     
     
         10 . The at least one memory device of  claim 9  wherein the translating is performed after the reordering. 
     
     
         11 . The at least one memory device of  claim 9  wherein the reordering is further based on reordering probabilities calculated for each of the reordered plurality of word strings. 
     
     
         12 . The at least one memory device of  claim 8  wherein the selecting is based on scores assigned to the translations of the plurality of word strings. 
     
     
         13 . The at least one memory device of  claim 12 , the method further comprising generating the assigned scores based on a log-linear paradigm. 
     
     
         14 . The at least one memory device of  claim 8 , the method further comprising building translation knowledge prior to the generating and selecting. 
     
     
         15 . A system comprising:
 a computer;   the computer configured for generating, by a computer, a plurality of word strings corresponding to a syntactic tree structure that represents a source word string in a natural source language; and   the computer further configured for selecting a preferred translation from translations of the plurality of word strings, the translations resulting from translating the plurality of word strings into a natural target language, the translating based on a plurality of alignment matrices that map sample source sentences in the natural source language to sample target sentences in the natural target language.   
     
     
         16 . The system of  claim 15 , the computer further configured for reordering, according to the alignment matrices, terms of the generated plurality of word strings with regard to a formal syntax. 
     
     
         17 . The system of  claim 16  wherein the translating is performed after the reordering. 
     
     
         18 . The system of  claim 16  wherein the reordering is further based on reordering probabilities calculated for each of the reordered plurality of word strings. 
     
     
         19 . The system of  claim 15  wherein the selecting is based on scores assigned to the translations of the plurality of word strings. 
     
     
         20 . The system of  claim 19 , the computer further configured for generating the assigned scores based on a log-linear paradigm.