# US 11,307,892 B2 · Machine-learning for state determination and prediction

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,307,892 B2
- Document ID: US11307892B2
- Status: Granted
- Kind: B2
- Application number: 16/020,206
- Filed: 2018-06-27
- Published: 2020-01-02
- Granted: 2022-04-19
- Earliest claimed: 2018-06-27
- Assignees: Optum Services (Ireland) Limited
- Applicants: Optum Services (Ireland) Limited
- Inventors: Damian Kelly; Dominik Dahlem; Rick A. Hamilton
- Web page: https://patentgrep.com/patent/US11307892B2

## CPC

- G06F 9/4881: {Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues}
- G06F 9/451: Execution arrangements for user interfaces
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/00: Machine learning
- G16H 20/10: relating to drugs or medications, e.g. for ensuring correct administration to patients
- G16H 40/63: for local operation
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

Without timely reminders, certain individuals may be highly unlikely to complete various tasks scheduled for completion. For example, medicinal therapy regimen patients often have difficulty in remembering to take their prescribed medications in accordance with care provider specified instructions. By providing a reminder prompt generation system that utilizes machine-learning and/or artificial intelligence to determine appropriate reminder times and modalities for providing automated-prompting notifications to users, the users' adherence to scheduled task completions can be significantly increased.

## Claims

### Claim 1

An apparatus comprising at least one processor and at least one memory including computer program code for one or more programs, the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following:
store a schedule comprising a task to be performed by a user;
provide a first automated-prompting notification to the user to perform the task;
determine that the user completed the task responsive to the first automated-prompting notification, wherein (a) a first state of the user and a first state of the user's environment are stored in association with the first automated-prompting notification, (b) the first state of the user, the first state of the user's environment, and the first automated-prompting notification are used to train a prompting model using machine learning, and (c) the prompting model outputs one or more automated-prompting notifications;
detect a first input indicative of a context of an environment of the user;
detect a first input indicative of a context of the user;
determine, based at least in part on the first input indicative of the context of the environment of the user and the first input indicative of the context of the user, a second state of the user and a second state of the user's environment;
determine, based at least in part on the second state of the user, the second state of the user's environment, and the prompting model, a second automated-prompting notification; and
provide the second automated-prompting notification.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the apparatus to perform at least the following:
determine that the user completed the task responsive to the second automated-prompting notification; and
train the prompting model using machine learning based at least in part on the second state of the user, the second state of the user's environment, and the second automated-prompting notification.

### Claim 3 (depends on 2)

The apparatus of claim 2, wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the apparatus to perform at least the following:
detect a second input indicative of a context of an environment of the user;
detect a second input indicative of a context of the user; and
determine, based at least in part on the second input indicative of the context of the environment of the user and the second input indicative of the context of the user, a future state of the user and a future state of the user's environment; and
determine, based at least in part on the future state of the user and the future state of the user's environment, a prompting model.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein the first input indicative of the context of the environment of the user is selected from the group consisting of an activation state of a device, a proximity to an object, a proximity to a device, and a sound.

### Claim 5 (depends on 1)

The apparatus of claim 1, wherein the first input indicative of the context of the user is selected from the group consisting of a movement of the user, an activity of the user, an action of the user, a gesticulation of the user, a vocal expression of the user, a facial expression of the user, a schedule of the user, a visual perception of the user, an auditory perception of the user, a tactile perception of the user, a cognitive load of the user, a social restriction of the user, an interaction of the user with a device, and interaction of the user with other users, an interaction of the user with a medical dispenser, an interaction of the user with services, and an inferred intention of the user.

### Claim 6 (depends on 1)

The apparatus of claim 1, wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the apparatus to perform at least the following:
control an output of a particular device by providing at least one control signal to the particular device.

### Claim 7 (depends on 6)

The apparatus of claim 6, wherein the prompting model is trained using the generative approach comprising at least one of Naïve Bayes, Linear Discriminant Analysis, Regularized Discriminant Analysis, Quadratic Discriminant Analysis, Gaussian Mixture Models, or Kernel Density Estimation.

### Claim 8 (depends on 1)

The apparatus of claim 1, wherein the prompting model is trained using a generative approach or a discriminative approach, and wherein the output of the prompting model predicts a likelihood of the user performing the task given a state of the user and a state of the user's environment.

### Claim 9 (depends on 1)

The apparatus of claim 1, wherein the prompting model is trained using the discriminative approach comprising at least one of Logistic Regression, Random Forests, Extreme Gradient Boosting, and Deep Neural Networks.

### Claim 10

A method for providing automated-prompting notifications for a user, the method comprising:
storing, via a non-transitory memory storage area, a schedule comprising a task to be performed by a user;
providing, via one or more processing devices, a first automated-prompting notification to the user to perform the task;
determining, via the one or more processing devices, that the user completed the task responsive to the first automated-prompting notification, wherein (a) a first state of the user and a first state of the user's environment are stored in association with the first automated-prompting notification, (b) the first state of the user, the first state of the user's environment, and the first automated-prompting notification are used to train a prompting model using machine learning, and (c) the prompting model outputs one or more automated-prompting notifications;
detecting, via the one or more processing devices, a first input indicative of a context of an environment of the user;
detecting, via the one or more processing devices, a first input indicative of a context of the user;
determining, via the one or more processing devices and based at least in part on the first input indicative of the context of the environment of the user and the first input indicative of the context of the user, a second state of the user and a second state of the user's environment;
determining, via the one or more processing devices and based at least in part on the second state of the user, the second state of the user's environment, and the prompting model, a second automated-prompting notification; and
providing, via the one or more processing devices, the second automated-prompting notification.

### Claim 11 (depends on 10)

The method of claim 10, further comprising:
determining that the user completed the task responsive to the second automated-prompting notification; and
training the prompting model using machine learning based at least in part on the second state of the user, the second state of the user's environment, and the second automated-prompting notification.

### Claim 12 (depends on 11)

The method of claim 11, further comprising:
detecting a second input indicative of a context of an environment of the user;
detecting a second input indicative of a context of the user; and
determining, based at least in part on the second input indicative of the context of the environment of the user and the second input indicative of the context of the user, a future state of the user and a future state of the user's environment; and
determining, based at least in part on the future state of the user and the future state of the user's environment, a prompting model.

### Claim 13 (depends on 10)

The method of claim 10, wherein the first input indicative of the context of the environment of the user is selected from the group consisting of an activation state of a device, a proximity to an object, a proximity to a device, and a sound.

### Claim 14 (depends on 10)

The method of claim 10, wherein the first input indicative of the context of the user is selected from the group consisting of a movement of the user, an activity of the user, an action of the user, a gesticulation of the user, a vocal expression of the user, a facial expression of the user, a schedule of the user, a visual perception of the user, an auditory perception of the user, a tactile perception of the user, a cognitive load of the user, a social restriction of the user, an interaction of the user with a device, and interaction of the user with other users, an interaction of the user with a medical dispenser, an interaction of the user with services, and an inferred intention of the user.

### Claim 15 (depends on 10)

The method of claim 10, further comprising: controlling an output of a particular device by providing at least one control signal to the particular device.

### Claim 16 (depends on 15)

The method of claim 15, wherein the prompting model is trained using the generative approach comprising at least one of Naïve Bayes, Linear Discriminant Analysis, Regularized Discriminant Analysis, Quadratic Discriminant Analysis, Gaussian Mixture Models, or Kernel Density Estimation.

### Claim 17 (depends on 10)

The method of claim 10, wherein the prompting model is trained using a generative approach or a discriminative approach, and wherein the output of the prompting model predicts a likelihood of the user performing the task given a state of the user and a state of the user's environment.

### Claim 18 (depends on 10)

The method of claim 10, wherein the prompting model is trained using the discriminative approach comprising at least one of Logistic Regression, Random Forests, Extreme Gradient Boosting, and Deep Neural Networks.

### Claim 19

A computer program product comprising a non-transitory computer readable medium having computer program instructions stored therein, the computer program instructions when executed by a processor, cause the processor to:
store a schedule comprising a task to be performed by a user;
provide a first automated-prompting notification to the user to perform the task;
determine that the user completed the task responsive to the first automated-prompting notification, wherein (a) a first state of the user and a first state of the user's environment are stored in association with the first automated-prompting notification, (b) the first state of the user, the first state of the user's environment, and the first automated-prompting notification are used to train a prompting model using machine learning, and (c) the prompting model outputs one or more automated-prompting notifications;
detect a first input indicative of a context of an environment of the user;
detect a first input indicative of a context of the user;
determine, based at least in part on the first input indicative of the context of the environment of the user and the first input indicative of the context of the user, a second state of the user and a second state of the user's environment;
determine, based at least in part on the second state of the user, the second state of the user's environment, and the prompting model, a second automated-prompting notification; and
provide the second automated-prompting notification.

### Claim 20 (depends on 19)

The computer program product of claim 19, wherein the computer program instructions further cause the processor to:
determine that the user completed the task responsive to the second automated-prompting notification; and
train the prompting model using machine learning based at least in part on the second state of the user, the second state of the user's environment, and the second automated-prompting notification.

### Claim 21 (depends on 20)

The computer program product of claim 20, wherein the computer program instructions further cause the processor to:
detect a second input indicative of a context of an environment of the user;
detect a second input indicative of a context of the user; and
determine, based at least in part on the second input indicative of the context of the environment of the user and the second input indicative of the context of the user, a future state of the user and a future state of the user's environment; and
determine, based at least in part on the future state of the user and the future state of the user's environment, a prompting model.

### Claim 22 (depends on 19)

The computer program product of claim 19, wherein the first input indicative of the context of the environment of the user is selected from the group consisting of an activation state of a device, a proximity to an object, a proximity to a device, and a sound.

### Claim 23 (depends on 19)

The computer program product of claim 19, wherein the first input indicative of the context of the user is selected from the group consisting of a movement of the user, an activity of the user, an action of the user, a gesticulation of the user, a vocal expression of the user, a facial expression of the user, a schedule of the user, a visual perception of the user, an auditory perception of the user, a tactile perception of the user, a cognitive load of the user, a social restriction of the user, an interaction of the user with a device, and interaction of the user with other users, an interaction of the user with a medical dispenser, an interaction of the user with services, and an inferred intention of the user.

### Claim 24 (depends on 19)

The computer program product of claim 19, wherein the computer program instructions further cause the processor to: control an output of a particular device by providing at least one control signal to the particular device.

### Claim 25 (depends on 24)

The computer program product of claim 24, wherein the prompting model is trained using the generative approach comprising at least one of Naïve Bayes, Linear Discriminant Analysis, Regularized Discriminant Analysis, Quadratic Discriminant Analysis, Gaussian Mixture Models, or Kernel Density Estimation.

### Claim 26 (depends on 19)

The computer program product of claim 19, wherein the prompting model is trained using a generative approach or a discriminative approach, and wherein the output of the prompting model predicts a likelihood of the user performing the task given a state of the user and a state of the user's environment.

### Claim 27 (depends on 19)

The computer program product of claim 19, wherein the prompting model is trained using the discriminative approach comprising at least one of Logistic Regression, Random Forests, Extreme Gradient Boosting, and Deep Neural Networks.