High confidence predicted profiles for online dating matching using continuously learning machine learning models trained for each user with physical and facial attributes

A device, comprising:

• a processor;
• and
• a memory that stores executable instructions that, when executed by the processor, facilitate performance of operations, comprising:
• training, by the device, a machine learning model with first feature vectors representative of potential first match entities identified by at least one user of the device as inputs to a selection process for matching entities to be matches for the at least one user of the device, the first feature vectors respectively comprising:
• first ratios applicable to facial first attributes of the potential first match entities extracted from first facial recognition of first images of the potential first match entities,
• first heights of the potential first match entities,
• first hair colors of the potential first match entities,
• first eye colors of the potential first match entities, or
• first biases attributable to the at least one user of the device applicable to the selection process for the matching entities determined by first natural language processing of first textual content attributable to the potential first match entities, resulting in a trained machine learning model, wherein the trained machine learning model is operable in a foreground execution space of the device, and wherein the trained machine learning model is further trained in a background processing space of the device while the trained machine learning model is usable for the selection process for the matching entities in the foreground execution space;
• updating the trained machine learning model comprising training the trained machine learning model further with second feature vectors representative of potential second match entities identified by the at least one user of the device as further inputs to the selection process for the matching entities, the second feature vectors comprising:
• second ratios applicable to second facial attributes of the potential second match entities extracted from second facial recognition of second images of the potential second match entities,
• second heights of the potential second match entities,
• second hair colors of the potential second match entities,
• second eye colors of the potential second match entities, or
• second biases attributable to the at least one user of the device further applicable to the selection process for the matching entities determined by second natural language processing of second textual content attributable to the at least one user of the device, resulting in a further trained machine learning model, wherein the updating the trained machine learning model comprises:
• swapping the trained machine learning model out of the foreground execution space, and swapping the further trained machine learning model into the foreground execution space, resulting in an untrained machine learning model being in the foreground execution space and the trained machine learning program being in the background processing space, and
• iteratively performing the updating and the swapping for additional feature vectors representative of additional potential match entities received by the device as additional inputs to the selection process for the matching entities, thereby continuously evolving the further trained machine learning model as a continuously evolved trained machine learning model applicable to the selection process for the matching entities;
• in response to a defined criterion being satisfied, injecting outlier potential match entities as supplemental inputs to the training of the trained machine learning model further with supplemental feature vectors representative of supplemental match entities, not identified by the at least one user of the device, as further inputs to the selection process for the matching entities, the supplemental feature vectors comprising:
• supplemental ratios applicable to supplemental facial attributes of a potential supplemental match entities extracted from third facial recognition of supplemental images of the potential second match entities,
• heights of the potential supplemental match entities,
• hair colors of the potential supplemental match entities, or
• eye colors of the potential supplemental match entities;
• for candidate profiles accessible to the selection process for the matching entities, applying the continuously evolved trained machine learning model to determine respective confidence values corresponding to the candidate profiles, wherein the respective confidence values represent respective likelihoods that the candidate profiles are part of the matching entities of the selection process to be the matches for the at least one user;
• and
• as a defined function of the respective confidence values, rendering at least profile images of profiles of the candidate profiles for the selection process.

A system, comprising:

• at least one processor, comprising:
• a first machine learning model in a foreground execution space;
• and
• a second machine learning model in a background processing space;
• and
• at least one memory that stores executable instructions that, when executed by the processor, facilitate performance of operations, comprising:
• training, by the system, the first and second machine learning models with at least one first feature vector representative of at least one potential first match entity identified by a user of the system as input to a selection process for at least one matching entity to be a match for the user, the at least one first feature vector comprising at least one of:
• at least one first ratio applicable to at least one facial first attribute of the at least one potential first match entity extracted from first facial recognition of at least one first image of the at least one potential first match entity,
• at least one first height of the at least one potential first match entity,
• at least one first hair color of the at least one potential first match entity,
• at least one first eye color of the at least one potential first match entity, or
• at least one first bias attributable to the user of the system applicable to the selection process for the at least one matching entity determined by first natural language processing of first textual content attributable to the at least one potential first match entity, resulting in a trained first machine learning model, wherein the second machine learning model is further trained while the trained first machine learning model is usable for the selection process for the at least one matching entity in the foreground execution space, resulting in a trained second machine learning model;
• updating the trained second machine learning model comprising training the trained second machine learning model further with at least one second feature vector representative of at least one potential second match entity identified by the user of the system as further input to the selection process for the at least one matching entity, the at least one second feature vector comprising at least one of:
• at least one second ratio applicable to at least one second facial attribute of the at least one potential second match entity extracted from second facial recognition of at least one second image of the at least one potential second match entity,
• at least one second height of the at least one potential second match entity,
• at least one second hair color of the at least one potential second match entity,
• at least one second eye color of the at least one potential second match entity, or
• at least one second bias attributable to the user of the system further applicable to the selection process for the at least one matching entity determined by second natural language processing of second textual content attributable to the user of the system, resulting in a further trained second machine learning model;
• in response to the updating of the trained second machine learning model, swapping the trained first machine learning model out of the foreground execution space, and swapping the further trained second machine learning model into the foreground execution space, resulting in the further trained second machine learning model being in the foreground execution space and the trained first machine learning model being in the background processing space;
• in response to a defined criterion being satisfied, injecting an outlier potential match entity as supplemental input to the training of the trained first machine learning model and the further trained second machine learning model further with a supplemental feature vector representative of a supplemental match entity, not identified by the user of the system, as further input to the selection process for the at least one matching entity, the supplemental feature vector comprising at least one of:
• a supplemental ratio applicable to a supplemental facial attribute of a potential supplemental match entity extracted from third facial recognition of a supplemental image of the at least one potential second match entity,
• a height of the potential supplemental match entity,
• a hair color of the potential supplemental match entity, or
• an eye color of the potential supplemental match entity;
• iteratively performing the updating, the swapping, and the injecting for additional feature vectors representative of additional potential match entities received by the system as additional input to the selection process for the at least one matching entity, thereby continuously evolving the trained first machine learning model and the further trained second machine learning model as a continuously evolved trained machine learning model applicable to the selection process for the at least one matching entity;
• for candidate profiles accessible to the selection process for the at least one matching entity, applying the continuously evolved trained machine learning model to determine respective confidence values corresponding to the candidate profiles, wherein the respective confidence values represent respective likelihoods that the candidate profiles are part of the at least one matching entity of the selection process to be the match for the user;
• and
• as a defined function of the respective confidence values, rendering at least a profile image of profiles of the candidate profiles for the selection process.