| CPC G06F 40/30 (2020.01) [G06Q 30/016 (2013.01); G06Q 50/26 (2013.01); G06N 5/04 (2013.01)] | 18 Claims |
|
1. A system for artificial intelligence directed controlled-environment facility resident support ticket response and/or action, the system comprising:
one or more controlled-environment facility media and/or communications terminals and/or a plurality of personal controlled-environment facility resident media and/or communications devices, each comprising at least one processor and a memory coupled to the at least one processor and storing program instructions thereupon, which when executed by the at least one processor, cause the respective controlled-environment facility media and/or communications terminal or personal controlled-environment facility resident media and/or communications device to:
provide access to a controlled-environment facility media and/or communications terminal or device support application program; and
accept input of a controlled-environment facility media and/or communications terminal or device support request, via the controlled-environment facility media and/or communications terminal or device support application program, from a controlled-environment facility resident accessing the respective controlled-environment facility media and/or communications terminal or personal controlled-environment facility resident media and/or communications device;
a controlled-environment facility media and/or communications device services platform comprising at least one processor and a memory coupled to the at least one processor and storing program instructions thereupon, which when executed by the at least one processor, cause the controlled-environment facility media and/or communications device services platform to:
train a machine learning natural language process model for controlled-environment facility resident support request response and/or action, employing examples of text comprising words, sentences, and paragraphs written by humans, to understand context of written text in the controlled-environment facility media and/or communications terminal or device support request;
receive the support request;
automatically direct the machine learning natural language process model to implement one or more phases of lexical analysis, parsing, semantic analysis, discourse integration, and pragmatic analysis to implement at least natural language process text-analytics including one or more of entity recognition, concept extraction, text classification, sentiment analysis, and relation extraction or parsing applied to text in the support request, to determine context of the text in the support request;
automatically direct the machine learning natural language process model to pass the support request through the machine learning natural language process model to determine a previously assigned weight of one or more words in the support request, employing the machine learning natural language process model determined context of text in the support request;
automatically direct the machine learning natural language process model to select one or more passages of predefined response text to compose text for a response to the support request, based, at least in part, on the machine learning natural language process model determined context of the support request, the machine learning natural language process model determined previously assigned weight of one or more words in the support request and one or more tags associated with the one or more words and the one or more passages of predefined text that observe rules of grammar;
automatically direct an artificial intelligence engine to, via a feedback loop, reiteratively review the response to confirm a correct response; and
automatically update the previously assigned weight to one or more words from the support request, based, at least in part, on whether the artificial intelligence engine reiterative review of the response confirms the response as the correct response; and
automatically update selected one or more passages of predefined response text and/or the one or more tags associated with the one or more words and the one or more passages of predefined text, based, at least in part, on whether the artificial intelligence engine reiterative review of the response confirms the response as the correct response;
automatically direct the machine learning natural language process model to personalize, using the machine learning natural language process model determined context of text of the support request, the response, based, at least in part, on identification of a resident; and
automatically send a resulting personalized response to the respective controlled-environment facility media and/or communications terminal or personal controlled-environment facility resident media and/or communications device.
|