Round trip system for parsing meaning
A computer-implemented method, a system, and a non-transitory computer-readable storage medium are described. In an embodiment, a computer-implemented method involves operating a natural language understanding process utilizing at least one of a Machine Learning Model and a Deep Learning Model that generates at least one of the following: entities, attributes, relationships and meaning, connected through at least one of a domain and a context in time, producing an operable meaning. Said natural language understanding process involves utilizing said entities, attributes, relationships and meaning as searchable data, presenting an input to be understood by a computing system and presented as a grammatically correct syntax based on said at least one of said domain and said context in time, and extracting at least one of an entity, attribute, relationship and meaning in association with to said grammatically correct syntax to be stored as semantic data knowledge.
1 . A computer-implemented method comprising a plurality of operations:
a natural language model containing an inference engine receives an input;
said input generates at least one user language model;
said natural language model derives a meaning from said input;
at least one of said inference engine and a domain knowledge module accesses a personal history of change for a source of said input;
said meaning of said input, and at least one meaning saved to said personal history of change that is relevant to said meaning of said input, are utilized to define a parsing of at least one specific meaning saved to a semantic database;
a round trip of processing of said at least one specific meaning within said semantic database is carried out by a plurality of processes, wherein said processes comprise active speech recognition, language modelling, information retrieval, and natural language understanding, which outputs an operable meaning;
said operable meaning is received by a functional state module;
said functional state module checks to see if it is possible to generate an output meaning based on an existing meaning of the functional state's defined relationships and conditions affected by said operable meaning from said user language model;
said semantic database processes a meaning of at least one of said domain knowledge module, a parsing engine module and a relationships module, plus the output meaning of said functional state module, to produce a combined meaning, and outputs said combined meaning as an operable meaning, wherein the active speech recognition runs for a first time period, wherein at an end of the first time period, the language modelling receives an ongoing output from the active speech recognition and continues for a same amount of time as the active speech recognition while the active speech recognition continues, and wherein after a second time period which occurs after the first time period but during the language modelling, the information retrieval begins and concludes with an outputting of the operable meaning, and
wherein said round trip is less than or equal to 2.6 picoseconds.
2 . The computer-implemented method of claim 1 , wherein said semantic database processes known meaning.
3 . The computer-implemented method of claim 1 , wherein said functional state module performs two additional tasks: verifying a state of said operable meaning, and producing the output meaning to an operator of said state of said operable meaning.
4 . The computer-implemented method of claim 1 , wherein the meaning of said user language model is defined by the meaning of a user's input at a point in time.
5 . The computer-implemented method of claim 1 , wherein the meaning of said user language model is defined by the meaning of said user's input according to a context of use.
6 . The computer-implemented method of claim 1 , wherein new information that is used to update said semantic database is tied to said personal history of change of a user who presents said input.
7 . The computer-implemented method of claim 1 , wherein information extraction is based on both context in time and a plurality of specific characteristics of a user who presents an input to said natural language model.
8 . The computer-implemented method of claim 1 , wherein an updating of said operable meaning by said functional state module is an ongoing process.
9 . The computer-implemented method of claim 1 , wherein extracted information is based on both context in time and a plurality of specific characteristics of a user.
10 . The computer-implemented method of claim 1 , wherein the information extraction matches correct information to a correct user at a context in time.
11 . The computer-implemented method of claim 1 , wherein the meaning of said user's input is saved to the said user's personal history of change and becomes part of known knowledge of said semantic database.
12 . The computer-implemented method of claim 1 , wherein said user language model is saved to a user's personal history of change.
13 . The computer-implemented method of claim 1 , wherein the language modelling comprises a plurality of operations of a context engine, a next word prediction module, and a grammar correction module.
14 . The computer-implemented method of claim 1 , wherein the information retrieval comprises a plurality of operations of the domain knowledge module, the parsing engine module, and the relationships module.