# US 12,002,587 B1 · Proxy model using mobile device data to provide health indicators > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,002,587 B1 - Document ID: US12002587B1 - Status: Granted - Kind: B1 - Application number: 15/880,043 - Filed: 2018-01-25 - Granted: 2024-06-04 - Earliest claimed: 2018-01-25 - Assignees: Blueowl, LLC - Applicants: Blueowl, LLC - Inventors: Kenneth J. Sanchez; Bennett Smith - Web page: https://patentgrep.com/patent/US12002587B1 ## CPC - G16H 50/30: for calculating health indices; for individual health risk assessment - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/045: Combinations of networks - G06N 3/08: Learning methods - G06N 3/084: Backpropagation, e.g. using gradient descent - G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems - G16H 50/50: for simulation or modelling of medical disorders - G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Systems and methods are provided for determining a health indicator, such as a life expectancy, of a target person using a proxy model comprising an artificial neural network, in lieu of more complex, costly, and/or invasive health assessment processes. The artificial neural network may be constructed from labeled training data comprising a plurality of activity data training sets, each set including personal activity metrics associated with activities performed by a respective training person as derived from sensor data obtained via one or more mobile electronic devices (e.g., a smartphone or a wearable device) of the respective training person. The trained artificial neural network may then obtain and process an activity data set associated with the target person to determine a health indicator of the target person. ## Claims ### Claim 1 A computer-implemented method for training an artificial neural network for health assessment, the method comprising: collecting labeled training data for the artificial neural network, the labeled training data including a plurality of activity data training sets corresponding to a plurality of training persons, wherein each activity data training set of the plurality of activity data training sets includes a respective plurality of personal activity metrics obtained or derived from one or more respective mobile electronic devices associated with a respective training person of the plurality of training persons, and the each activity data training set is labeled with a respective known health indicator assigned to the respective training person by a health assessment process; training the artificial neural network with the labeled training data by at least identifying, from the respective plurality of personal activity metrics in the plurality of activity data training sets, one or more significant metrics having a significant predictive effect upon the respective known health indicator, and one or more insignificant metrics not having a significant predictive effect upon the respective known health indicator, wherein the artificial neural network is trained when the artificial neural network determines a respective output for the each activity data training set that matches the respective known health indicator associated with the each activity data training set; obtaining an activity data set corresponding to a target person, the activity data set including a plurality of personal activity metrics obtained or derived from one or more mobile electronic devices associated with the target person; receiving, by the artificial neural network, as trained, the activity data set; processing, by the artificial neural network, as trained, the activity data set corresponding to the target person; determining, by the artificial neural network, as trained, a health indicator of the target person based upon processing the activity data set corresponding to the target person; and outputting, by the artificial neural network, as trained, the health indicator of the target person, as determined; wherein the obtaining the activity data set includes: activating one or more sensors to obtain sensor data associated with the one or more significant metrics, and deactivating the one or more sensors to not obtain sensor data associated with the one or more insignificant metrics. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, wherein the health assessment process is at least partially manual. ### Claim 3 (depends on 2) The computer-implemented method of claim 2, wherein the health assessment process relies upon medical history information associated with the respective training person. ### Claim 4 (depends on 1) The computer-implemented method of claim 1, wherein training the artificial neural network includes comparing a health indicator determined by the artificial neural network processing an activity data training set of the plurality of activity data training sets corresponding to a particular training person from among the plurality of training persons to a known health indicator assigned to the particular training person by the health assessment process. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, wherein the artificial neural network is a recurrent neural network having a plurality of hidden layers, and wherein training the artificial neural network includes determining respective weights of at least one of: (i) one or more personal activity metrics provided as inputs to the artificial neural network, or (ii) one or more outputs generated by one or more of the plurality of hidden layers. ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein: one or more of the plurality of activity data training sets are obtained via respective applications installed on respective mobile electronic devices associated with respective training persons of the plurality of training persons; and the respective applications are configured to obtain sensor data at one or more predetermined frequencies of time, via activation of one or more respective sensors of the respective mobile electronic devices. ### Claim 7 (depends on 1) The computer-implemented method of claim 1, wherein: the activity data set corresponding to the target person is obtained via an application installed on at least one device of the one or more mobile electronic devices associated with the target person; and the application is configured to obtain sensor data at a predetermined frequency of time, via activation of one or more sensors of the at least one device. ### Claim 8 A computer system configured to train an artificial neural network for health assessment, the computer system comprising: one or more processors; and one or more memories storing computer-executable instructions that, when executed by the one or more processors, cause the computing system to: collect labeled training data for the artificial neural network, the labeled training data including a plurality of activity data training sets corresponding to a plurality of training persons, wherein each activity data training set of the plurality of activity data training sets includes a respective plurality of personal activity metrics obtained or derived from one or more respective mobile electronic devices associated with a respective training person of the plurality of training persons, and the each activity data training set is labeled with a respective known health indicator assigned to the respective training person by a health assessment process; train the artificial neural network with the labeled training data by at least identifying, from the respective plurality of personal activity metrics in the plurality of activity data training sets, one or more significant metrics having a significant predictive effect upon the respective known health indicator, and one or more insignificant metrics not having a significant predictive effect upon the respective known health indicator, wherein the artificial neural network is trained when the artificial neural network determines a respective output for the each activity data training set that matches the respective known health indicator associated with the each activity data training set; obtain an activity data set corresponding to a target person, the activity data set including a plurality of personal activity metrics obtained or derived from one or more mobile electronic devices associated with the target person; receive, by the artificial neural network, as trained, the activity data set; process, by the artificial neural network, as trained, the activity data set corresponding to the target person; determine, by the artificial neural network, as trained, a health indicator of the target person based upon processing the activity data set corresponding to the target person; and output, by the artificial neural network, as trained, the health indicator of the target person, as determined; wherein to obtain the activity data set includes: to activate one or more sensors to obtain sensor data associated with the one or more significant metrics, and to deactivate the one or more sensors to not obtain sensor data associated with the one or more insignificant metrics. ### Claim 9 (depends on 8) The computer system of claim 8, wherein the health assessment process is at least partially manual. ### Claim 10 (depends on 8) The computer system of claim 8, wherein the health assessment process relies upon medical history information associated with the respective training person. ### Claim 11 (depends on 8) The computer system of claim 8, wherein training the artificial neural network includes comparing a health indicator determined by the artificial neural network processing an activity data training set of the plurality of activity data training sets corresponding to a particular training person from among the plurality of training persons to a known health indicator assigned to the particular training person by the health assessment process. ### Claim 12 (depends on 8) The computer system of claim 8, wherein the artificial neural network is a recurrent neural network having a plurality of hidden layers, and wherein training the artificial neural network includes determining respective weights of at least one of: (i) one or more personal activity metrics provided as inputs to the artificial neural network, or (ii) one or more outputs generated by one or more of the plurality of hidden layers. ### Claim 13 (depends on 8) The computer system of claim 8, wherein: one or more of the plurality of activity data training sets are obtained via respective applications installed on respective mobile electronic devices associated with respective training persons of the plurality of training persons; and the respective applications are configured to obtain sensor data at one or more predetermined frequencies of time, via activation of one or more respective sensors of the respective mobile electronic devices. ### Claim 14 (depends on 8) The computer system of claim 8, wherein: the activity data set corresponding to the target person is obtained via an application installed on at least one device of the one or more mobile electronic devices associated with the target person; and the application is configured to obtain sensor data at a predetermined frequency of time, via activation of one or more sensors of the at least one device. ### Claim 15 A non-transitory, computer-readable medium for training an artificial neural network for health assessment, the non-transitory, computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to: collect labeled training data for the artificial neural network, the labeled training data including a plurality of activity data training sets corresponding to a plurality of training persons, wherein each activity data training set of the plurality of activity data training sets includes a respective plurality of personal activity metrics obtained or derived from one or more respective mobile electronic devices associated with a respective training person of the plurality of training persons, and the each activity data training set is labeled with a respective known health indicator assigned to the respective training person by a health assessment process; train the artificial neural network with the labeled training data by at least identifying, from the respective plurality of personal activity metrics in the plurality of activity data training sets, one or more significant metrics having a significant predictive effect upon the respective known health indicator, and one or more insignificant metrics not having a significant predictive effect upon the respective known health indicator, wherein the artificial neural network is trained when the artificial neural network determines a respective output for the each activity data training set that matches the respective known health indicator associated with the each activity data training set; obtain an activity data set corresponding to a target person, the activity data set including a plurality of personal activity metrics obtained or derived from one or more mobile electronic devices associated with the target person; receive, by the artificial neural network, as trained, the activity data set; process, by the artificial neural network, as trained, the activity data set corresponding to the target person; determine, by the artificial neural network, as trained, a health indicator of the target person based upon processing the activity data set corresponding to the target person; and output, by the artificial neural network, as trained, the health indicator of the target person, as determined; wherein to obtain the activity data set includes: to activate one or more sensors to obtain sensor data associated with the one or more significant metrics, and to deactivate the one or more sensors to not obtain sensor data associated with the one or more insignificant metrics. ### Claim 16 (depends on 15) The non-transitory, computer-readable medium of claim 15, wherein the instructions further cause the one or more processors to train the artificial neural network by comparing a health indicator determined by the artificial neural network processing an activity data training set of the plurality of activity data training sets corresponding to a particular training person from among the plurality of training persons to a known health indicator assigned to the particular training person by the health assessment process. ### Claim 17 (depends on 15) The non-transitory, computer-readable medium of claim 15, wherein the artificial neural network is a recurrent neural network having a plurality of hidden layers, and wherein the instructions further cause the one or more processors to train the artificial neural network by determining respective weights of at least one of: (i) one or more personal activity metrics provided as inputs to the artificial neural network, or (ii) one or more outputs generated by one or more of the plurality of hidden layers. ### Claim 18 (depends on 15) The non-transitory, computer-readable medium of claim 15, wherein: the activity data set corresponding to the target person is obtained via an application installed on at least one device of the one or more mobile electronic devices associated with the target person; and the application is configured to obtain sensor data at a predetermined frequency of time, via activation of one or more sensors of the at least one device. ### Claim 19 (depends on 15) The non-transitory, computer-readable medium of claim 15, wherein one or more of: the health assessment process is at least partially manual; or the health assessment process relies upon medical history information associated with the respective training person. ### Claim 20 (depends on 15) The non-transitory, computer-readable medium of claim 15, wherein: one or more of the plurality of activity data training sets are obtained via respective applications installed on respective mobile electronic devices associated with respective training persons of the plurality of training persons; and the respective applications are configured to obtain sensor data at one or more predetermined frequencies of time, via activation of one or more respective sensors of the respective mobile electronic devices.