Humans use common language to seamlessly exchange knowldege. Artificial General Intelligence (AGI) will be possible when AI systems have a common language.

COMFIED tackles the AGI challenge in a new way – through a 7×7 decision template. Requests containing a non-visual grid layout will be sent to an online API or open source JavaScript library to get an intelligent decision response.

This simple but versatile design could be used to add intelligent features to consumer devices, generate music and many other creative and productive applications. Instead of exchanging large volumes of complex machine learning models, the car and music generator would cross-transfer their AI knowledge by simply exchange their 7×7 grid templates. These knowledge templates frame the problem in small modules that can be cascaded and quickly tested for bias and efficacy.

QUÉ is a demonstration of an application using this concept packaged as a puzzle game. It runs on a 10-layer decision engine that adapts with different scenarios, while complying with rules and available choices. The next step is to convert this game play engine into a hosted service and JavaScript library, and avail it for use in apps, websites and IoT devices.