Artificial intelligence server

1. An artificial intelligence server, comprising:

a communication interface configured to receive a plurality of user dictionaries included in a plurality of user profiles; and

a processor including a profile classifier, a ranking calculator, a cloud user dictionary generator, and an updater;

the processor configured to:

classify the plurality of user profiles by a same user profile using the profile classifier,

calculate rankings of a plurality of keywords included in the user dictionary corresponding to the same user profile using the ranking calculator,

generate a cloud user dictionary including keywords having a specific ranking or higher according to the rankings of the plurality of keywords using the cloud user dictionary generator, and

train a natural language processing model using the cloud user dictionary through the updater,

wherein the specific ranking is a predetermined ranking value having a specific score, and

all keywords included in the cloud user dictionary have a value greater than or equal to the predetermined ranking value,

wherein the cloud user dictionary includes a first keyword, a first combination keyword, and a second combination keyword corresponding a specific profile among the plurality of user profiles,

the first keyword is a keyword which is not recognized in the natural language processing model,

the first combination keyword is a combination of the first keyword which is not recognized in the natural language processing model and a second keyword recognized in the natural language processing model, and

the second combination keyword is a combination of the second keyword and a third keyword recognized in the natural language processing model, and

wherein the natural language processing model is trained by using the first keyword, the first combination keyword, and the second combination keyword corresponding to the specific profile among the plurality user profiles to generate a trained natural language processing model, and

the trained natural language processing model is transmitted to an external device of a user corresponding to the specific profile through the communication interface.

2. The artificial intelligence server of claim 1 , wherein the processor is configured to calculate the rankings of the plurality of keywords, by giving a high score to a keyword which is frequently detected in the user dictionary.

3. The artificial intelligence server of claim 1 , wherein the user dictionary includes the plurality of keywords which are words not recognized in the natural language processing model.

4. The artificial intelligence server of claim 3 , wherein the plurality of keywords includes a third word which is a combination of a first word and a second word recognized in the natural language processing model.

5. The artificial intelligence server of claim 3 , wherein the plurality of keywords includes a third word which is a combination of a first word which is not recognized in the natural language processing model and a second word recognized in the natural language processing model.

6. A method of operating an artificial intelligence server having a communication interface and a processor including a profile classifier, a ranking calculator, a cloud user dictionary generator, and an updater, the method comprising:

receiving a plurality of user dictionaries included in a plurality of user profiles through the communication interface;

classifying the plurality of user profiles by a same user profile through the profile classifier;

calculating rankings of a plurality of keywords included in the user dictionary corresponding to the same user profile through the ranking calculator;

generating a cloud user dictionary including keywords having a specific ranking or higher according to the rankings of the plurality of keywords through the cloud user dictionary generator; and

training a natural language processing model using the generated cloud user dictionary through the updater,

wherein the specific ranking is a predetermined ranking value having a specific score, and

all keywords included in the cloud user dictionary have a value greater than or equal to the predetermined ranking value,

wherein the cloud user dictionary includes a first keyword, a first combination keyword, and a second combination keyword corresponding a specific profile among the plurality of user profiles,

the first keyword is a keyword which is not recognized in the natural language processing model,

the first combination keyword is a combination of the first keyword which is not recognized in the natural language processing model and a second keyword recognized in the natural language processing model, and

the second combination keyword is a combination of the second keyword and a third keyword recognized in the natural language processing model, and

wherein the natural language processing model is trained by using the first keyword, the first combination keyword, and the second combination keyword corresponding to the specific profile among the plurality of user profiles to generate a trained natural language processing model, and

the trained natural language processing model is transmitted to an external device of a user corresponding to the specific profile through the communication interface.

7. The method of claim 6 , wherein the calculating of the rankings of the plurality of keywords included in the user dictionary includes calculating the rankings of the plurality of keywords, by giving a high score to a keyword which is frequently detected in the user dictionary.

8. The method of claim 6 , wherein the user dictionary includes the plurality of keywords which are words not recognized in the natural language processing model.

9. The method of claim 8 , wherein the plurality of keywords includes a third word which is a combination of a first word and a second word recognized in the natural language processing model.

10. The method of claim 8 , wherein the plurality of keywords includes a third word which is a combination of a first word which is not recognized in the natural language processing model and a second word recognized in the natural language processing model.

11. An artificial intelligence server, comprising:

a communication interface configured to receive a plurality of user dictionaries included in a plurality of user profiles corresponding to a plurality of users from a plurality of external devices, each of the plurality of external devices corresponding to a different user among the plurality of users; and

a processor including a profile classifier, a ranking calculator, a cloud user dictionary generator, and an updater;

the processor configured to:

collect keywords from at least two different users corresponding to at least two user profiles,

classify the plurality of user profiles based on the at least two user profiles among the plurality of user profiles using the profile classifier,

calculate rankings of the keywords from the at least two different users corresponding to the at least two user profiles using the ranking calculator,

generate a cloud user dictionary including the keywords collected from the at least two different users corresponding to the at least two user profiles and having a specific ranking or higher according to the rankings of the keywords using the cloud user dictionary generator, and

train a natural language processing model using the cloud user dictionary through the updater,

wherein the specific ranking is a predetermined ranking value having a specific score, and

all keywords included in the cloud user dictionary have a value greater than or equal to the predetermined ranking value,

wherein the cloud user dictionary includes a first keyword, a first combination keyword, and a second combination keyword corresponding to the at least two user profiles,

the first keyword is a keyword which is not recognized in the natural language processing model,

the first combination keyword is a combination of the first keyword which is not recognized in the natural language processing model and a second keyword recognized in the natural language processing model, and

the second combination keyword is a combination of the second keyword and a third keyword recognized in the natural language processing model, and

wherein the natural language processing model is trained by using the first keyword, the first combination keyword, and the second combination keyword corresponding to the specific profile among the plurality user profiles to generate a trained natural language processing model, and

the trained natural language processing model is transmitted to an external device corresponding to at least one of the two different users through the communication interface.