Artificial intelligence based health coaching based on ketone levels of participants
First Claim
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.
6 Assignments
0 Petitions
Accused Products
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.
85 Citations
16 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. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
-
Specification