Methods and systems of interchanging code-mixed words and uni-language words

Patent metadata

Field	Value
Publication number	US-11151326-B2
Application number	US-201916365714-A
Country	US
Kind code	B2
Filing date	Mar 27, 2019
Priority date	Feb 13, 2019
Publication date	Oct 19, 2021
Grant date	Oct 19, 2021

Abstract

A method and a systems of interchanging code-mixed words and uni-language words are described. In an embodiment, the method may include identifying two or more portions of a target word, wherein the target word is one of a code-mixed word and a uni-language word. The method may further include determining one or more equivalent words corresponding to at least one portion of the two or more portions, wherein the one or more equivalent words are in at least one predetermined target language. The method may further include selecting at least one equivalent word from the one or more equivalent words, based on a context of the target word within a sentence comprising the target word. The method may further include replacing the at least one portion of the two or more portions of the target word with the selected at least one equivalent word.

First claim

What is claimed is:

1. A method of interchanging code-mixed words and uni-language words, the method comprising:
identifying, by a translation device, two or more portions of a target word, wherein the target word is one of a code-mixed word and a uni-language word;
determining, by the translation device, one or more equivalent words corresponding to at least one portion of the two or more portions, wherein the one or more equivalent words are in at least one predetermined target language, wherein determining the one or more equivalent words comprises identifying a derivative word for a portion of the at least one portion to be replaced, and wherein the derivative word is in a native language associated with the portion;
selecting, by the translation device, at least one equivalent word from the one or more equivalent words, based on a context of the target word within a sentence comprising the target word; and
replacing, by the translation device, the at least one portion of the two or more portions of the target word with the selected at least one equivalent word.

2. The method of claim 1 , wherein at least one remaining portion of the two or more portions is an ending phoneme corresponding to a verb.

3. The method of claim 2 , wherein the at least one predetermined target language is determined based on the at least one portion that is an ending phoneme.

4. The method of claim 1 , wherein the two or more portions are identified using a bidirectional Long Short Term Memory (LSTM).

5. The method of claim 1 , wherein the at least one predetermined target language corresponds to a user preference.

6. The method of claim 1 , wherein selecting the at least one equivalent word comprises identifying the at least one equivalent word associated with the derivative word.

7. The method of claim 6 further comprising:
computing a first probability score for each of the at least one equivalent word, based on an associated usage probability, wherein the associated usage probability of an equivalent word of the at least one equivalent word is derived based on linguistic rules of the at least one predetermined target language; and
computing a second probability score for each of the at least one equivalent word, based on an associated transition probability, wherein the associated transition probability of the equivalent word of the at least one equivalent word is derived based on a placement of the equivalent word within the sentence comprising the target word.

8. The method of claim 7 further comprising:
determining a combined probability score for each of the at least one equivalent word based on the associated first probability score and the second probability score; and
selecting an equivalent word from the at least one equivalent word based on the combined probability score determined for each of the at least one equivalent word.

9. The method of claim 8 further comprising:
receiving an alternate word suggestion for the equivalent word from a user; and
training the translation device based on the alternate word suggested by the user.

10. A translation device for interchanging code-mixed words and uni-language words, the translation device comprising:
a processor; and
a memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to:
identify two or more portions of a target word, wherein the target word is one of a code-mixed word and a uni-language word;
determine one or more equivalent words corresponding to at least one portion of the two or more portions, wherein the one or more equivalent words are in at least one predetermined target language, wherein determining the one or more equivalent words comprises identifying a derivative word for a portion of the at least one portion to be replaced, and wherein the derivative word is in a native language associated with the portion;
select at least one equivalent word from the one or more equivalent words, based on a context of the target word within a sentence comprising the target word; and
replace the at least one portion of the two or more portions of the target word with the selected at least one equivalent word.

11. The translation device of claim 10 , wherein at least one remaining portion of the two or more portions is an ending phoneme corresponding to a verb, and wherein the at least one predetermined target language is determined based on the at least one portion that is an ending phoneme.

12. The translation device of claim 10 , wherein the two or more portions are identified using a bidirectional Long Short Term Memory (LSTM).

13. The translation device of claim 10 , wherein the at least one predetermined target language corresponds to a user preference.

14. The translation device of claim 10 , wherein selecting the at least one equivalent word comprises identifying the at least one equivalent word associated with the derivative word.

15. The translation device of claim 14 , wherein the processor instructions further cause the processor to:
compute a first probability score for each of the at least one equivalent word, based on an associated usage probability, wherein the associated usage probability of an equivalent word of the at least one equivalent word is derived based on linguistic rules of the at least one predetermined target language; and
compute a second probability score for each of the at least one equivalent word, based on an associated transition probability, wherein the associated transition probability of the equivalent word of the at least one equivalent word is derived based on a placement of the equivalent word within the sentence comprising the target word.

16. The translation device of claim 15 , wherein the processor instructions further cause the processor to:
determine a combined probability score for each of the at least one equivalent word based on the associated first probability score and the second probability score; and
select an equivalent word from the at least one equivalent word based on the combined probability score determined for each of the at least one equivalent word.

17. The translation device of claim 16 , wherein the processor instructions further cause the processor to:
receive an alternate word suggestion for the equivalent word from a user; and
train the translation device based on the alternate word suggested by the user.

18. A non-transitory computer-readable storage medium having stored thereon, a set of computer-executable instructions causing a computer comprising one or more processors to perform steps comprising:
identifying two or more portions of a target word, wherein the target word is one of a code-mixed word and a uni-language word;
determining one or more equivalent words corresponding to at least one portion of the two or more portions, wherein the one or more equivalent words are in at least one predetermined target language, wherein determining the one or more equivalent words comprises identifying a derivative word for a portion of the at least one portion to be replaced, and wherein the derivative word is in a native language associated with the portion;
selecting at least one equivalent word from the one or more equivalent words, based on a context of the target word within a sentence comprising the target word; and
replacing the at least one portion of the two or more portions of the target word with the selected at least one equivalent word.

Assignees

• Wipro Ltd

Inventors

• Ramachandra Iyer Manjunath
• Chaitanya Villari Boby

Classifications

• G06F40/253

  Grammatical analysis; Style critique · CPC title

• G06F40/263 Primary

  Language identification · CPC title

• G06F40/44

  Statistical methods, e.g. probability models · CPC title

• G06F40/279

  Recognition of textual entities · CPC title

• G06F40/30 Primary

  Semantic analysis · CPC title