Granted Patent B2

US 12,639,394 · App. 18/609,563 · Granted May 26, 2026

Inventors: Barbara Sue Smith (Toronto, CA); Daniel J. Sullivan (Toronto, CA)

Assignee: The Strategic Coach Inc.

G06F17/11G06N20/00

Abstract

Claims (58)

1 . An apparatus for optimization using dynamically variable local parameters, wherein the apparatus comprises:

at least a processor; and

a memory communicatively connected to the at least a processor, wherein the memory contains instructions configuring the at least a processor to:

access a first dataset corresponding to a first phenomenon comprising a user behavior and a second dataset corresponding to a second phenomenon comprising an environmental factor, wherein accessing the second dataset further comprises:

determining the environmental factor using a sensor incorporated in the at least a processor, wherein the environmental factor comprises a location of a user;

process, using an encoder, the first dataset, wherein the processed first dataset comprises a reduced-dimensionality representation compared to a known input space, wherein the encoder comprises a neural network configured to jointly optimize across input data and output a plurality of parameters corresponding to variational distributions, wherein the reduced-dimensionality representation is used by the at least a processor to dynamically adjust parameters in real time, and wherein a representation comprises at least a feature, wherein the at least a feature comprises at least a user engagement trend;

identify a dependency relationship between at least one first data cluster of the processed first dataset and at least one second data cluster of the second dataset;

modify an attribute set of the at least one second data cluster as a function of the dependency relationship, wherein modifying the attribute set comprises generating an objective function;

optimize a target data cluster as a function of the processed first dataset and the modified attribute set; and

modify the at least one second data cluster as a function of the target data cluster, wherein modifying the at least one second data cluster comprises determining a user specific behavioral prompt as a function of the processed first dataset using a generative machine-learning model by:

sanitizing training data to eliminate noise, wherein the training data comprises the processed first dataset, wherein sanitizing the training data comprises:

determining that at least one training data entry of the training data has a signal to noise ratio below a threshold value; and

removing the at least one training data entry from the training data to create sanitized training data;

training the generative machine-learning model on the sanitized training data until the generative machine-learning model satisfies a convergence test, wherein sanitizing of the training data accelerates convergence of the generative machine-learning model, wherein the generative machine-learning model comprises a generative adversarial network and at least a discriminator, wherein the at least a discriminator comprises a supervised machine-learning model, and wherein the generative adversarial network is configured to generate hypothetical processed first dataset examples as a function of feedback from the at least a discriminator;

generating the user specific behavioral prompt using the generative machine-learning model;

updating the sanitized training data with user engagement data corresponding to the user specific behavioral prompt; and

retraining the generative machine-learning model using the updated sanitized training data.

2 . The apparatus of claim 1 , further configured to generate the second dataset using a machine-learning process.

3 . The apparatus of claim 1 , wherein modifying the attribute set includes adding an optimization constraint.

4 . The apparatus of claim 1 , wherein modifying the attribute set includes removing an optimization constraint.

5 . The apparatus of claim 1 , wherein generating the objective function further comprises:

selecting, using a machine-learning model, a variable set; and

generating the objective function using the selected variable set.

6 . The apparatus of claim 1 , wherein generating the objective function further comprises:

receiving a variable set;

initializing a coefficient set corresponding to the variable set; and

tuning the coefficient set using a machine-learning algorithm.

7 . The apparatus of claim 1 , wherein identifying the dependency relationship comprises determining a strength of a correlation between the at least one first data cluster of the processed first dataset and the at least one second data cluster.

8 . The apparatus of claim 1 , wherein the apparatus is further configured to predict an impact of changes in the at least one second data cluster on the optimization of the target data cluster.

9 . The apparatus of claim 8 , wherein predicting the impact of changes in the at least one second data cluster further comprises predicting an impact of changes in the at least one first data cluster of the processed first dataset as a function of a hypothetical adjustments to the at least one second data cluster.

10 . A method for optimization using dynamically variable local parameters, the method comprising:

accessing, using at least a processor, a first dataset corresponding to a first phenomenon comprising a user behavior and a second dataset corresponding to a second phenomenon comprising an environmental factor, wherein accessing the second dataset further comprises:

determining the environmental factor using a sensor incorporated in the at least a processor, wherein the environmental factor comprises a location of a user;

processing, using an encoder, the first dataset, wherein the processed first dataset comprises a reduced-dimensionality representation compared to a known input space, wherein the encoder comprises a neural network configured to jointly optimize across input data and output a plurality of parameters corresponding to variational distributions, wherein the reduced-dimensionality representation is used by the at least a processor to dynamically adjust parameters in real time, and wherein a representation comprises at least a feature, wherein the at least a feature comprises at least a user engagement trend;

identifying, using the at least a processor, a dependency relationship between at least one first data cluster of the processed first dataset and at least one second data cluster of the second dataset;

modifying, using the at least a processor, an attribute set, as a function of the dependency relationship, wherein modifying the attribute set comprises generating an objective function;

optimizing, using the at least a processor, a target data cluster, as a function of the processed first dataset and the modified attribute set; and

modifying, using the at least a processor, the at least one second data cluster as a function of the target data cluster, wherein modifying the at least one second data cluster comprises determining a user specific behavioral prompt as a function of the processed first dataset using a generative machine-learning model by:

sanitizing training data to eliminate noise, wherein the training data comprises the processed first dataset, wherein sanitizing the training data comprises:

determining that at least one training data entry of the training data has a signal to noise ratio below a threshold value; and

removing the at least one training data entry from the training data to create sanitized training data;

training the generative machine-learning model on the sanitized training data until the generative machine-learning model satisfies a convergence test, wherein sanitizing of the training data accelerates convergence of the generative machine-learning model, wherein the generative machine-learning model comprises a generative adversarial network and at least a discriminator, wherein the at least a discriminator comprises a supervised machine-learning model, and wherein the generative adversarial network is configured to generate hypothetical processed first dataset examples as a function of feedback from the at least a discriminator;

generating the user specific behavioral prompt using the generative machine-learning model;

updating the sanitized training data with user engagement data corresponding to the user specific behavioral prompt; and

retraining the generative machine-learning model using the updated sanitized training data.

11 . The method of claim 10 , further comprises generating, using the at least a processor, the second dataset using a machine-learning process.

12 . The method of claim 10 , wherein modifying the attribute set includes adding an optimization constraint.

13 . The method of claim 10 , wherein modifying the attribute set includes removing an optimization constraint.

14 . The method of claim 10 , wherein generating the objective function further comprises:

selecting, using a machine-learning model, a variable set; and

generating the objective function using the selected variable set.

15 . The method of claim 1 , wherein generating the objective function further comprises:

receiving a variable set;

initializing a coefficient set corresponding to the variable set; and

tuning the coefficient set using a machine-learning algorithm.

16 . The method of claim 10 , wherein identifying the dependency relationship comprises determining a strength of a correlation between the at least one first data cluster of the processed first dataset and the at least one second data cluster.

17 . The method of claim 10 , further comprises predicting, using the at least a processor, an impact of changes in the at least one second data cluster on the optimization of the target data cluster.

18 . The method of claim 17 , wherein predicting the impact of changes in the at least one second data cluster further comprises predicting an impact of changes in the at least one first data cluster of the processed first dataset as a function of a hypothetical adjustments to the at least one second data cluster.

Assignments (1)

Continuity (1)

References Cited (9)