System and method for query authorization and response generation using machine learning
Systems, methods, and computer-readable storage media for responding to a query using a neural network and natural language processing. If necessary, the system can request disambiguation, then parse the query using a trained machine-learning classifier, resulting in at least one of an identified subject or an identified domain of the text query. The system can determine if the user is authorized to retrieve answers to the query and, if so, retrieve factual data associated with the query. The system can then retrieve a response template, and fill in the template with the retrieved facts. The system can then determine, by executing a machine comprehension model on the filled response template, a probable readability token, portion of text, of at least a portion of the filled response template and, upon identifying that the probable readability is above a threshold, reply to the text query with the at least a portion of the filled response template.
1. A method comprising:
identifying, at a processor, an identity of a user;
receiving a text query of the user;
identifying, via the processor executing a classifier, at least one named entity within the text query;
retrieving, via the processor from a personnel database, a plurality of identities associated with the at least one named entity;
receiving, from the user, a disambiguation input identifying a selected entity from the plurality of identities;
parsing, via the processor executing a trained machine-learning classifier, the text query, resulting in at least one of an identified subject or an identified domain of the text query;
authorizing the user, via the processor and based on the identity of the user and the at least one of the identified subject or the identified domain, to access a restricted database containing restricted data associated with the at least one of the identified subject or the identified domain;
retrieving, via the processor, a response template specific to the at least one of the identified subject or the identified domain;
filling, via the processor accessing the restricted database, the response template with at least a portion of the restricted data associated with the at least one of the identified subject or the identified domain, resulting in a filled response template;
removing portions of the filled response template which are not associated with the at least one of the identified subject or the identified domain, resulting in at least a portion of the filled response template;
determining, via the processor executing a machine comprehension model on the filled response template, a probable readability token of the at least a portion of the filled response template; and
upon identifying that the probable readability is above a threshold, replying to the text query with the at least a portion of the filled response template.
2. The method of claim 1 , wherein the classifier is a neural network is pretrained to identify proper nouns within the text query.
3. The method of claim 1 , wherein the trained machine-learning classifier iteratively modifies code executed by the processor upon receiving a threshold number of text queries.
4. The method of claim 3 , wherein the trained machine-learning classifier uses logistic regression between iterations to identify which aspects of the code to modify.
5. The method of claim 1 , wherein the response template includes an identifier associated with the portion of the restricted data, and further comprising:
matching the identifier with a key database to retrieve data corresponding to the identifier, and
filling the response template with the data.
6. The method of claim 1 , wherein the text query of the user is:
received as a speech query; and
converted, by the processor executing a speech-to-text conversion, the speech query into the text query.
7. The method of claim 1 , wherein the at least one named entity within the text query comprises at least one of a legal formatted name, a given name, a family name, and a nickname.
8. A system comprising:
a processor; and
a non-transitory computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising:
identifying an identity of a user;
receiving a text query of the user;
identifying, by executing a classifier, at least one named entity within the text query;
retrieving, from a personnel database, a plurality of identities associated with the at least one named entity;
receiving, from the user, a disambiguation input identifying a selected entity from the plurality of identities;
parsing, by executing a trained machine-learning classifier, the text query, resulting in at least one of an identified subject or an identified domain of the text query;
authorizing the user, based on the identity of the user and the at least one of the identified subject or the identified domain, to access a restricted database containing restricted data associated with the at least one of the identified subject or the identified domain;
retrieving a response template specific to the at least one of the identified subject or the identified domain;
filling, by accessing the restricted database, the response template with at least a portion of the restricted data associated with the at least one of the identified subject or the identified domain, resulting in a filled response template;
removing portions of the filled response template which are not associated with the at least one of the identified subject or the identified domain, resulting in at least a portion of the filled response template;
determining, by executing a machine comprehension model on the filled response template, a probable readability token of the at least a portion of the filled response template; and
upon identifying that the probable readability is above a threshold, replying to the text query with the at least a portion of the filled response template.
9. The system of claim 8 , wherein the classifier is a neural network is pretrained to identify proper nouns within the text query.
10. The system of claim 8 , wherein the trained machine-learning classifier iteratively modifies code executed by the processor upon receiving a threshold number of text queries.
11. The system of claim 10 , wherein the trained machine-learning classifier uses logistic regression between iterations to identify which aspects of the code to modify.
12. The system of claim 8 , wherein the response template includes an identifier associated with the portion of the restricted data, and further comprising:
matching the identifier with key database to retrieve data corresponding to the identifier, and
filling the response template with the data.
13. The system of claim 8 , wherein the at least one named entity within the text query comprises at least one of a legal formatted name, a given name, a family name, and a nickname.
14. A non-transitory computer-readable storage medium having instructions stored which, when executed by a computing device, cause the computing device to perform operations comprising:
identifying an identity of a user;
receiving a text query of the user;
identifying, by executing a classifier, at least one named entity within the text query;
retrieving, from a personnel database, a plurality of identities associated with the at least one named entity;
receiving, from the user, a disambiguation input identifying a selected entity from the plurality of identities;
parsing, by executing a trained machine-learning classifier, the text query, resulting in at least one of an identified subject or an identified domain of the text query;
authorizing the user, based on the identity of the user and the at least one of the identified subject or the identified domain, to access a restricted database containing restricted data associated with the at least one of the identified subject or the identified domain;
retrieving a response template specific to the at least one of the identified subject or the identified domain;
filling, by accessing the restricted database, the response template with at least a portion of the restricted data associated with the at least one of the identified subject or the identified domain, resulting in a filled response template;
removing portions of the filled response template which are not associated with the at least one of the identified subject or the identified domain, resulting in at least a portion of the filled response template;
determining, by executing a machine comprehension model on the filled response template, a probable readability of the at least a portion of the filled response template; and
upon identifying that the probable readability is above a threshold, replying to the text query with the at least a portion of the filled response template.
15. The non-transitory computer-readable storage medium of claim 14 , wherein the classifier is a neural network is pretrained to identify proper nouns within the text query.
16. The non-transitory computer-readable storage medium of claim 14 , wherein the trained machine-learning classifier iteratively modifies code executed by the processor upon receiving a threshold number of text queries.
17. The non-transitory computer-readable storage medium of claim 16 , wherein the trained machine-learning classifier uses logistic regression between iterations to identify which aspects of the code to modify.