Abstract

This document describes a framework for a personalization engine to suggest custom configuration parameters to each user, by creating a ‘digital twin’ of the user. The proposed framework implements artificial intelligence and/or machine learning principles to predict the preferred user experience and probable behavior of the user and observe actual behavior. The model is designed to be self-improving by comparing the model output (predicted behavior) to ground-truth (actual behavior). As the machine learning model is trained on the user’s preferences, historical behavior and digital actions, the model then provides real-time, custom-tailored personalization suggestions to the user. The model is either cached and pre-executed or runs continuously at super-human speeds. The unified system can be efficiently scaled and coordinated, increasing the integration of user personalization across many applications. In some instances, the model assists in cooperative multitasking, which involves continuing one task across multiple applications. The model may also create and provide options and decision support based on user activity and preferences in all applications from which it receives data.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Download

Share

COinS