Systems and methods to infer user behavior
Inventors
Levchuk, Georgiy • Schurr, Nathan • Hering, Darby • Zakin, Mitch
Assignees
Publication Number
US-11651285-B1
Publication Date
2023-05-16
Expiration Date
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Abstract
Methods to infer user behavior are disclosed comprising a process of predefining one or more activities for a user application, providing a processor based device and user interface configured to operate with the device to support the user with tasks using the user application, the user application communicating an intent message to a transformative power management (TPM) application and the TPM configured to define and output an instruction for the application given the intention and the predefined activities. In some embodiments, the methods are implemented on a processor based device. In some embodiments, the systems and methods apply pattern recognition algorithms and pattern learning algorithms to manage the power allocation to power consuming devices.
Core Innovation
Methods to infer user behavior are disclosed comprising a process of predefining one or more activities for a user application, providing a processor based device and user interface configured to operate with the device to support the user with tasks using the user application, the user application communicating an intent message to a transformative power management (TPM) application and the TPM configured to define and output an instruction for the application given the intention and the predefined activities. In some embodiments, the methods are implemented on a processor based device. In some embodiments, the systems and methods apply pattern recognition algorithms and pattern learning algorithms to manage the power allocation to power consuming devices.
Users of technology today balance many opposing challenges including increasingly sophisticated apps, multiple apps running at once, individual user differences, and reliance on users to proactively configure power management options; this mismatch between handheld power needs and the resources available to meet these needs creates a need for advanced energy management. To address this need, the Transformative Power Management (TPM) system applies pattern recognition algorithms to forecast user specific resource allocations based on observable data and inferring user behavior and activities such as tasks, and in some embodiments pattern learning algorithms are also used to enhance the ability of the system to correlate or infer parameters.
Claims Coverage
This patent includes three independent claims. The following lists the main inventive features extracted from those independent claims.
Method to modify an application activity
Predefining a plurality of application activities to be executed by an application; ordering a subset of the plurality of application activities as a planned sequence of application activities given an application task with a planned next application activity to follow a current application activity; receiving an intent message from the application wherein the intent message defines one of the plurality of application activities as the current application activity; inferring a user behavior from the intent message by automatically determining a probability of the user behavior from an action with a pattern recognition algorithm where the action comprises at least a first and second action, the pattern recognition algorithm comprises a first set of pattern recognition algorithm parameters for the first action and a second set of pattern recognition algorithm parameters for the second action, and updating the second set of pattern recognition algorithm parameters according to a pattern learning algorithm; determining a selected next application activity from the plurality of application activities given the user behavior and the plurality of application activities; and outputting an instruction to modify the planned sequence of application activities whereby the selected next application activity follows the current application activity.
Method to modify a sequence of application activities
Predefining one or more application activity for an application and a planned sequence of application activities; providing a user interface to the application; receiving an intent message from the user interface; determining a probability of a user behavior from the intent message with a pattern recognition algorithm where the intent message can comprise a first intent message representing a first action and a second intent message representing a second action and the pattern recognition algorithm comprises a first set of parameters for the first action and a second set of parameters for the second action and updating the second set according to a pattern learning algorithm; defining a forecasted application activity from the one or more application activity given the intent message, the one or more application activity and the user behavior; determining a modified sequence of application activities given the forecasted application activity; and outputting an instruction to modify the planned sequence of application activities to the modified sequence of application activities.
Behavior inference system to infer a user behavior and modify a subsequent user behavior
One or more processors and one or more memory elements with instructions to: predefine one or more application activity and a planned sequence of application activities; provide a user interface and receive an intent message; infer a current user behavior from the intent message with a user model and infer a mission state from the intent message with a mission model; determine a forecasted application activity with an application use model given the current user behavior and the mission state; infer a forecasted user behavior and a modified user behavior with the user model given the forecasted application activity; determine a modified application activity and a modified sequence of application activities with the application use model given the modified user behavior; output an instruction to the user interface to modify a subsequent user behavior from the forecasted user behavior to the modified user behavior whereby the subsequent user behavior modifies the planned sequence of application activities to the modified sequence of application activities; and wherein inferring the current user behavior comprises automatically determining a probability of the current user behavior from an action with a pattern recognition algorithm that comprises first and second sets of parameters for at least first and second actions and updating the second set according to a pattern learning algorithm.
The independent claims focus on (1) methods that receive intent messages, infer user behavior probabilistically with pattern recognition and learning, forecast or select next application activities, and output instructions to modify planned sequences, and (2) a system implementing these functions via user, mission, and application use models to modify application activity sequences and associated power allocations.
Stated Advantages
Apply pattern recognition algorithms and/or pattern learning algorithms to enhance the management of resources.
Manage the power allocation to power consuming devices.
Forecast power demand based on a mission profile and user behavior and dynamically prioritize and allocate resources accordingly for optimal power use.
Provide user-individualized power resource management for devices running multiple apps.
Extend mission duration, simplify logistics, and increase device usability and consumer satisfaction.
Enable 'energy-aware' mission or application planning to minimize energy usage across an entire mission or application set without loss of effectiveness.
Power modes simplify the power optimization process and allow TPM to specify optimal power modes for applications.
Adaptively and proactively trade off current energy utilization using knowledge of current and future power demands and availability and find globally optimal app run modes across applications.
Documented Applications
Manage the power consumption of electronic devices.
Manage the power consumption of handheld devices.
Manage the power consumption of military equipment.
Manage building power allocation and electrical grid power allocation.
Allocate other types of resources including personnel resource allocation, allocation of attention such as in a command and control environment, allocation of budgets, or allocation of computing resources.
Operate as middleware for a processor based system such as the Android operating system to provide user-individualized power resource management for devices running multiple apps.
Serve as a control center for power management on devices to dynamically set priorities for optimization of power resource allocation.
Interested in licensing this patent?