Language-understanding training database action pair augmentation using bidirectional translation
View Patent ↗Embodiments of the present invention relate to a method and system for augmenting a training database of an automated language-understanding system. In one embodiment, a training example in a first language is received from the training database. The first language-training example translated to a second language output. The second language output is translated to a first variant of the first language-training example. An action pair including the first variant of the first language-training example and an action command associated with the first language-training example is stored in an augmented training database.
1. A method for augmenting an initial training database of an automated language-understanding system, comprising:
receiving a training example in a first language from the initial training database, wherein the first language training example is from an action pair comprising the first langage training example and an action command associated with the first language training example;
translating the first language training example to a second language output;
translating the second language output to a first variant of the first language training example; and
storing an augmented action pair including the first variant of the first language-training example and the action command associated with the first language-training example in an augmented training database.
2. The method of claim 1 , further comprising:
translating the second language output to a third language output;
translating the third language output to a second variant of the first language training example; and
outputting the second variant of the first language-training example to an output manager.
3. The method of claim 2 , further comprising:
storing an augmented action pair including the second variant of the first language-training example and the action command associated with the first language-training example in the augmented training database.
4. The method of claim 1 , wherein the action command is a computer executable command.
5. The method of claim 1 , wherein the first language-training example is in at least one of: English, German, French, Portuguese, Italian, Japanese, Korean, Chinese, Russian, Greek, Armenian, or Hindi.
6. The method of claim 1 , wherein the second language output is in at least one of: English, German, French, Portuguese, Italian, Japanese, Korean, Chinese, Russian, Greek, Armenian, or Hindi.
7. The method of claim 1 , wherein the first variant is in at least one of: English, German, French, Portuguese, Italian, Japanese, Korean, Chinese, Russian, Greek, Armenian, or Hindi.
8. A system for augmenting an initial training database of an automated language-understanding system, comprising:
a first language-translation unit configured to translate a first language-training example to a second language output, wherein the first language-training example is from an action pair comprising the first language-training example and an action command associated with the first language-training example;
a second language-translation unit configured to translate the second language output to a first variant of the first language training example; and
a database configured to store an augmented action pair including the first variant of the first language-training example and the action command associated with the first language-training example.
9. The system of claim 8 , further comprising:
a third language-translation unit configured to translate the second language output to a third language output; and
a fourth language-translation unit configured to translate the third language output to a second variant of the first language-training example.
10. The system of claim 9 , further comprising:
an output manager configured to output the second variant of the first language-training example.
11. The system of claim 9 , wherein the database is configured to store the second variant of the first language-training example and the action command associated with the first language-training example.
12. A machine-readable medium having stored thereon a plurality of executable instructions, the plurality of instructions comprising instructions to:
receive a training example in a first language from the initial training database, wherein the first language training example is from an action pair comprising the first language training example and an action command associated with the first language training example;
translate the first language training example to a second language output;
translate the second language output to a first variant of the first language training example; and
store an augmented action pair including the first variant of the first language-training example and the action command associated with the first language-training example in an augmented training database.
13. The machine-readable medium of claim 12 having stored thereon additional executable instructions, the additional instructions comprising instructions to:
translate the second language output to a third language output;
translate the third language output to a second variant of the first language training example; and
output the second variant of the first language-training example to an output manager.
14. The machine-readable medium of claim 13 having stored thereon additional executable instructions, the additional instructions comprising instructions to:
store an augmented action pair including the second variant of the first language-training example and the action command associated with the first language-training example in the augmented training database.
15. A system for augmenting an initial training database of an automated language-understanding system, comprising:
a language-translation unit configured to translate a first language-training example to a second language output and configured to translate the second language output to a first variant of the first language training example, wherein the first language-training example is from an action pair comprising the first language-training example and an action command associated with the first language-training example; and
a database configured to store an augmented action pair including the first variant of the first language-training example and the action command associated with the first language-training example.
16. The system of claim 15 , wherein the language translation unit further configured to translate the second language output to a third language output and to translate the third language output to a second variant of the first language-training example.
17. The system of claim 16 , further comprising:
an output manager configured to output the second variant of the first language-training example.
18. The system of claim 16 , wherein the database is configured to store the second variant of the first language-training example and the action command associated with the first language-training example.