System And Method For Language Independent Iterative Learning Mechanism For Nlp Tasks

Claims:WE CLAIM:

1. A method of language independent iterative learning mechanism for Natural Language Processing (NLP) tasks, the method comprising: identifying, by an NLP device, at least one NLP feature associated with a set of words within a sentence for an NLP task; identifying, by the NLP device, a linkage relationship between each subset of two adjacent words within the set of words; creating a pattern associated with the sentence for the NLP task, based on the at least one NLP feature associated with the set of words and the linkage relationship between each subset of two adjacent words; computing a confidence score corresponding to the pattern, based on a comparison within a trained dataset; assigning a pattern category to the pattern, based on the confidence score and a predefined threshold score, wherein the pattern category corresponds to the NLP task; and executing the NLP task based on the assigned pattern category.

2. The method of claim 1, wherein the pattern comprises a sequence of the set of words, wherein each subset of two adjacent words in the sequence are interlinked based on an associated linkage relationship.

3. The method of claim 1, wherein the trained dataset is created for the NLP task and comprises at least one of: a plurality of unique patterns, wherein each of the plurality of unique patterns corresponds to the at least one NLP feature derived for the NLP task; and a plurality of unique tuples derived from the plurality of unique patterns.

4. The method of claim 1, wherein the confidence score is derived based on a weighted average of a first confidence score and a second confidence score, wherein, the first confidence score is a pattern confidence score determined for the pattern, and wherein the pattern confidence score is derived based on a percentage match of the pattern with one of the plurality of unique patterns, wherein the pattern confidence score is independent of the language associated with the sentence; and wherein, the second confidence score is a tuple confidence score determined for each of the set of words in the sentence, and wherein the tuple confidence score is derived based on semantic similarity with at least one of the plurality of unique tuples within the trained dataset, wherein the tuple confidence score is dependent of the language associated with the sentence.

5. The method of claim 1, wherein the pattern category is assigned to the pattern when the confidence score is greater than or equal to the predefined threshold score.

6. The method of claim 1 further comprising identifying the pattern as a new pattern, when the confidence score is below the predefined threshold score.

7. A system for language independent iterative learning mechanism for Natural Language Processing (NLP) tasks, the system 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 at least one NLP feature associated with a set of words within a sentence for an NLP task; identify a linkage relationship between each subset of two adjacent words within the set of words; create a pattern associated with the sentence for the NLP task, based on the at least one NLP feature associated with the set of words and the linkage relationship between each subset of two adjacent words; compute a confidence score corresponding to the pattern, based on a comparison within a trained dataset; assign a pattern category to the pattern, based on the confidence score and a predefined threshold score, wherein the pattern category corresponds to the NLP task; and execute the NLP task based on the assigned pattern category.

8. The system of claim 7, wherein the pattern comprises a sequence of the set of words, wherein each subset of two adjacent words in the sequence are interlinked based on an associated linkage relationship.

9. The system of claim 7, wherein the trained dataset is created for the NLP task and comprises at least one of: a plurality of unique patterns, wherein each of the plurality of unique patterns corresponds to the at least one NLP feature derived for the NLP task; and a plurality of unique tuples derived from the plurality of unique patterns.

10. The system of claim 7, wherein the confidence score is derived based on a weighted average of a first confidence score and a second confidence score, wherein, the first confidence score is a pattern confidence score determined for the pattern, and wherein the pattern confidence score is derived based on a percentage match of the pattern with one of the plurality of unique patterns, wherein the pattern confidence score is independent of the language associated with the sentence; and wherein, the second confidence score is a tuple confidence score determined for each of the set of words in the sentence, and wherein the tuple confidence score is derived based on semantic similarity with at least one of the plurality of unique tuples within the trained dataset, wherein the tuple confidence score is dependent of the language associated with the sentence.

Dated this 15th day of February, 2019

Madhusudan S.T Of K&S Partner Agent for the Applicant IN/PA-1297 , Description:TECHNICAL FIELD This disclosure relates generally to Natural Language Processing (NLP) tasks, and more particularly to system and method for language independent iterative learning mechanism for NLP tasks.