Automatically varying system clocks to simulate test environments for application triggers generated using machine learning

Inventors

Bajaj, Apoorv • Gao, Yinxiang • Manchanda, Ritesh • Qiu, Chuanhan

Assignees

Click Therapeutics Inc

Abstract

Provided herein are systems and methods for simulating testing environments for digital therapeutic applications to address conditions over spans of time. A computing system may select a trigger from a plurality of triggers of an event scheduler for addressing a condition. The computing system may set a clock to a timestamp of the trigger, wherein the clock is accessible to the service and to the user device. The computing system may cause the service to transmit a message for addressing the condition to the user device, responsive to the timestamp of the clock corresponding to the timestamp of the trigger. The computing system may identify that the message is communicated between the service and the user device in accordance with the trigger. The computing system may determine that the trigger of the event scheduler is validated, responsive to identifying that the message is communicated.

Core Innovation

The invention simulates testing environments for digital therapeutic applications to address conditions over spans of time by using an event scheduler whose triggers identify timestamps at which at least one of a service or a user device is to communicate a corresponding message for addressing the condition. A test management service selects the trigger and modifies a clock on at least one of the service or the user device to advance in time over a current time or over a timestamp of the trigger prior to an actual time corresponding to the timestamp.

The system causes the service and/or the user device to communicate the message for addressing the condition responsive to the clock corresponding to the timestamp of the trigger, and identifies that the message is communicated between the service and the user device in accordance with the trigger. The trigger validator validates each scheduler trigger based on whether the triggered communication satisfies trigger criteria, using transmission and receipt transaction logs and/or response time windows, and determines an overall validated or not-validated scheduler state.

The invention further supports failure handling by determining triggers based on whether the expected communications occur, including re-cause/retry message transmission and transaction-log-based outputs. An event scheduler may generate triggers dynamically using a machine-learning model, and triggers may include criteria beyond timestamps; message presentation can include SMS message, MMS message, and in-app message, including while medication is concurrent.

Claims Coverage

The document contains two independent claims, each covering a core workflow of event-scheduler-driven, clock-advanced message communication for digital therapeutic testing, together with validation of whether each trigger was satisfied. Across the independent claims, five main inventive features are repeatedly required: selecting timestamp-based scheduler triggers, advancing a service and/or user-device clock, communicating condition-addressing messages responsive to the trigger timestamp, identifying that the communication occurred in accordance with the trigger, and determining that the trigger is validated based on that identification.

Both independent claims cover simulating digital therapeutic testing environments over spans of time by selecting timestamp-based event-scheduler triggers, advancing a clock on a service and/or user device to those timestamps, communicating condition-addressing messages responsive to the trigger, identifying that communication occurred in accordance with the trigger, and determining that the trigger is validated based on that communication.

Stated Advantages

Documented Applications