Artificial intelligence based health coaching based on ketone levels of participants

• US 10,068,494 B2
• Filed: 05/05/2017
• Issued: 09/04/2018
• Est. Priority Date: 10/14/2016
• Status: Active Grant

• Current Assignee: Invoy Holdings, Inc.
• Original Assignee: Invoy Holdings, LLC
• Inventors: Ahmad, Lubna, Ahmad, Salman
• Primary Examiner(s): Hull, James
• Application Number:
  US15/588,414
• Publication Number:
  US 20180108272A1
• Time in Patent Office: 487 Days
• Field of Search: None
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G09B 19/00 Teaching not covered by other main groups of this subclass teaching or practice apparatus for gun-aiming or gun-laying F41G3/26
• CPC Class Codes: G09B 19/0092 Nutrition
• CPC Class Codes: G09B 5/02 with visual presentation of the material to be studied, e.g. using film strip
• CPC Class Codes: G16H 20/60 relating to nutrition control, e.g. diets
• CPC Class Codes: G16H 20/70 relating to mental therapies, e.g. psychological therapy or autogenous training
• CPC Class Codes: G16H 40/63 for local operation
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients

Abstract

A system is disclosed that uses profiles of users, including monitored ketone levels of the users, to assess effectiveness levels of health programs (such as weight loss programs) assigned to the users, and to select health program modifications for the users. The system may use a machine learning (artificial intelligence) algorithm to adaptively learn how to classify users and to select messaging and behavioral modifications for the users. For example, in some embodiments the system classifies the users and provides associated health program recommendations using a computer model trained with expert-classified user data records. As another example, a set of rules may be used to generate the health program recommendations and related messaging, and the set of rules may automatically be modified over time based on feedback data reflective of health program effectiveness levels produced by such rules. In some embodiments the system includes a mobile application that runs on mobile devices of users and communicates wirelessly with breath analysis devices of the users. The mobile application may also communicate with a server-based system that generates the health program recommendations.

Claims

1. A system capable of using artificial intelligence to provide health coaching based on breath ketone levels of users, the system comprising:

a plurality of portable breath analysis devices, each breath analysis device comprising a ketone sensor capable of measuring ketone levels in breath samples of users to generate ketone measurements of the users, the ketone measurements reflective of effectiveness levels of current health programs assigned to the users, each breath analysis device comprising a wireless transceiver capable of wirelessly transmitting the ketone measurements of the users; and

a computing system that hosts an automated health coaching system, the automated health coaching system configured to use a machine learning process to classify the users based, at least partly, on data records of the users, the data records including the ketone measurements of the users and including other profile data of the users, the automated health coaching system further configured to use at least the classifications to select health program modifications, including diet modifications, for particular users, and to output an indication of the selected health program modifications for display to the respective users via a user interface, the computing system comprising one or more physical servers;

wherein the computing system is programmed with executable instructions to use a trained model to classify the users based on the data records of the users, the trained model comprising (1) a feature extractor that extracts features from the data records of the users, the features including features based on the ketone measurements and other profile data of the users, and (2) a classifier that classifies the users using a set of weights that specify amounts of weight to apply to particular extracted features, the weights learned by applying a machine learning algorithm to classified user data records, wherein the machine learning algorithm comprises a neural network algorithm, a Support Vector Machine algorithm, a Probabilistic Graphic Model algorithm, or a Decision Tree model algorithm.

2. The system of claim 1, wherein the automated health coaching system is configured to assign to a user a classification that represents an effectiveness of a current health program assigned to the user.

3. The system of claim 1, wherein the automated health coaching system is configured to assign to a user a classification representing a rate at which the user'"'"'s ketone measurements increase in response to placement of the user on a particular diet.

4. The system of claim 1, wherein the automated health coaching system is configured to assign multiple concurrent classifications to a user.

5. The system of claim 1, wherein the automated health coaching system comprises a data repository of rules, including a rule that specifies an action to be performed in response to a change in a classification of a user.

6. The system of claim 5, further comprising an adaptive refinement component that adaptively modifies the data repository of rules based on feedback data reflective of effectiveness levels of the rules on health conditions of the users.

7. The system of claim 1, wherein the automated health coaching system is programmed to use a collaborative filtering algorithm to select health programs to recommend to users, the collaborative filtering algorithm configured to select, for a target user, at least one health program that has produced effective results for users identified as being similar to the target user.

8. The system of claim 1, further comprising a mobile application that runs on mobile devices of the users, the mobile application configured to enable a mobile device of a user to (1) communicate wirelessly with a breath analysis device of the user, and (2) communicate with the computing system, the mobile application further configured to display to the user health program modifications selected by the automated health coaching system for the user.

9. The system of claim 8, wherein the mobile application is configured to prompt the user for information regarding an effectiveness of a current health program, and the automated health coaching system is programmed to use the information to classify the user.

10. The system of claim 8, further comprising a wireless scale, the wireless scale comprising a wireless transceiver and being configured to report weight measurements to the mobile application.

11. The system of claim 10, wherein the system is programmed to use at least weight measurements generated with the wireless scale, in combination with ketone measurements generated by a breath analysis device, to classify a user in terms of whether the user is burning fat.

12. The system of claim 1, wherein the computing system is configured to adaptively adjust the weights based on monitored health program results resulting from the user classifications and the selected health program modifications.

13. The system of claim 1, wherein the classified user data records are human-expert-classified user data records.

14. The system of claim 1, wherein the classifier comprises sub-classifiers, and is capable of classifying a user such that the user is concurrently classified into multiple classifications.

15. The system of claim 1, wherein the trained model includes a plurality of classifiers, including a first classifier for generating a classifier used to select a first type of health program modification and a second classifier used to select a second type of health program modification.

16. The system of claim 1, wherein the feature extractor is configured to generate a feature that is based on a plurality, but less than all, ketone measurements included in a data record of a user.