Targeted generation of messages for digital therapeutics using generative transformer models

Inventors

Morse, William • Li, Qingzhe

Assignees

Click Therapeutics Inc

Abstract

Aspects of the present disclosure are directed to systems, methods, and computer readable media for generating messages targeted to address conditions in users. A computing system may identify (i) a condition of a user to be addressed and (ii) one or more parameters defining messages to be presented via an application towards achieving an endpoint associated with the condition for the user. The computing system may generate a prompt using the condition and the one or more parameters in accordance with a template. The computing system may apply the prompt to a generative model to output a message identifying at least one activity toward achieving the endpoint. The generative model may be trained using a corpus. The computing system may provide, the message for presentation to prompt the user to perform the activity via the application towards achieving the endpoint to address the condition.

Core Innovation

The document describes a digital therapeutics messaging service that provides messages targeted at addressing conditions in users. A computing system identifies a condition of a user and one or more parameters defining messages presented via an application toward achieving an endpoint associated with the condition. The system generates an input using the condition and the one or more parameters in accordance with a template, and provides the input to a generative transformer model to produce message data for presentation via the application.

The generative transformer model is established by identifying a plurality of corpora for training the model. For each corpus, the system provides at least a portion of the respective input to the generative transformer model to generate a respective output, compares a first distribution of tokens in the input and a second distribution of tokens in the output, and updates one or more of the plurality of weights in the generative transformer model based on the comparison.

Based on providing the input to the generative transformer model, the system generates data for a message identifying at least one activity toward achieving the endpoint. The service provides, for presentation, the message identifying the at least one activity to be performed by the user via the application toward achieving the endpoint to address the condition.

Claims Coverage

The document provides two independent claims, a method claim and a system claim, each built around condition-parameter identification, template-based input generation, generative transformer model training using token distribution comparisons with weight updates, and generation of messages identifying at least one activity toward an endpoint for the user. Dependent claims further refine delivery modality, candidate message selection, user behavior, user preferences, user profile information, multi-modal content including visual content, delivery timing via a delivery model, and response-driven updating.

Across the independent claims, the core coverage is a targeted digital-therapeutics messaging workflow that combines condition and parameter identification, template-based input generation, a generative transformer model trained using token distribution comparisons with weight updates, and the generation and presentation of message data identifying at least one activity toward an endpoint associated with addressing the condition.

Stated Advantages

Documented Applications