Machine language translation with transfer mappings having varying context
AssigneeMicrosoft Corporation
InventorsArul A. Menezes, Stephen D. Richardson
Life of this patent
Estimating the term…
Abstract
A computer-implemented machine translation system translates text from a first language to a second language. The system includes a plurality of mappings, each mapping indicative of associating a dependency structure of the first language with a dependency structure of the second language, wherein at least some of the mappings correspond to dependency structures of the first language having varying context with some common elements, and associated dependency structures of the second language to the dependency structures of the first language. A module receives input text in a first language and outputs output text in a second language based on accessing the plurality of mappings.
Claims19
A hardware computer readable storage media having information and instructions thereon for a computer-implemented machine translation system to translate text from a first language to a second language, the information and instructions comprising:
- a plurality of mappings, each mapping indicative of associating a dependency structure of the first language with a dependency structure of the second language, wherein at least some of the mappings correspond to dependency structures of the first language having varying context with some common elements, wherein first and second dependency structures of the first language have at least one common word but the second dependency structure includes an additional word providing additional context and associated dependency structures of the second language to the dependency structures of the first language also having varying context with some common elements, wherein third and fourth dependency structure of the second language are associated with the first and second dependency structures of the first language, respectively, and have at least one common word but the fourth dependency structure includes a further word providing additional context;
- and
- a module that receives input text in a first language having the second dependency and outputs output text having the fourth dependency structure in a second language based on accessing the plurality of mappings.
A machine translation system to translate text from a first language to a second language, comprising:
- memory storing a plurality of mappings, each mapping indicative of associating a dependency structure of the first language with a dependency structure of the second language, wherein at least some of the mappings correspond to dependency structures of the first language having varying context with some common elements, and associated dependency structures of the second language to the dependency structures of the first language, and wherein at least some of the mappings are indicative of overlapping corresponding dependency structures wherein a first mapping includes corresponding dependency structures in the first and second languages that are made up only of related lemma and wherein a second mapping includes corresponding dependency structures in the first and second languages that have at least a pair of related lemmas present in the first mapping and additional related elements not present in the first mapping, each of the additional elements not being a specific lemma;
- a processor configured to access the memory, receive input text in a first language and output text in a second language based on the overlapping corresponding dependency structures.
A computer implemented method to translate text from a first language to a second language, comprising:
- receiving input text in a first language;
- using a processor to access a plurality of mappings, each mapping indicative of associating a dependency structure of the first language with a dependency structure of the second language, wherein at least some of the mappings correspond to dependency structures of the first language having varying context with some common elements, and associated dependency structures of the second language to the dependency structures of the first language;
- identifying at least one mapping for translating the input text using additional information related to the mappings, wherein the additional information includes at least one of information indicative of a size of each dependency structure of the first language;
- information indicative of an extent of a complete alignment of the dependency structures of the first language originating from a larger dependency structure;
- and information indicative of a frequency the dependency structure occurred in training data;
- and
- outputting text in the second language based on using said at least one mapping.
Dependencies are inferred from the claim text.