For hiearchical decomposition deep reinforcement learning for an artificial intelligence model

A method to apply Reinforcement Learning for an Artificial Intelligence model for subsequent deployment of that AI model, comprising:

• applying a hierarchical-decomposition reinforcement learning technique to train one or more AI objects as concept nodes composed in a hierarchical graph incorporated into the AI model,
• using the hierarchical-decomposition reinforcement learning technique to hierarchically decompose a complex task into multiple smaller, individual sub-tasks making up the complex task, where one or more of the individual sub-tasks, correspond to its own concept node in the hierarchical graph, and initially train the AI objects on the individual sub-tasks and then train on how the individual sub-tasks need to interact with each other in the complex task in order to deliver an end solution to the complex task;
• using information supplied from a user interface to apply the hierarchical-decomposition reinforcement learning technique to train one or more AI objects;
• cooperating with one or more data sources to obtain data for training and to conduct the training of the two or more AI objects corresponding to concept nodes in the hierarchical graph, in parallel at the same time;
• and
• wherein decomposing the complex task, enables use of reward functions focused for solving each individual sub-task and then one or more reward functions focused for the end solution of the complex task, as well as enabling conducting the training of the AI objects corresponding to the individual sub-tasks in the complex task, in parallel, where a combined parallel training and the use of reward functions focused for solving each individual sub-task speed up an overall training duration for the complex task on one or more computing platforms, and subsequent deployment of a resulting AI model that is trained, compared to an end-to-end training with a single algorithm for all of the AI objects incorporated into the AI model.

An artificial intelligence engine, comprising:

• two or more modules configured to cooperate with each other in order to apply a hierarchical-decomposition reinforcement learning technique to train one or more AI objects as concept nodes composed into a hierarchical graph incorporated into an AI model, where individual sub-tasks of a decomposed task correspond to its own concept node in the hierarchical graph, where one or more of the individual sub-tasks of the decomposed task incorporated into its own hierarchical graph of concept nodes are initially trained on how to complete their individual sub-task and then an integrator node is trained to choose which of its children nodes is most appropriate for solving its individual sub-task at that moment, to deliver an end solution to the complex task, where during training of the individual sub-tasks of the decomposed task the AI engine is configured to use i) reward functions focused for solving each individual sub-task and then ii) a separate one or more reward functions focused for solving the end solution of the complex task with all of the individual sub-tasks, where the AI engine is configured to conduct the training of a first individual sub-task of the decomposed task incorporated into its own hierarchical graph of concept nodes in parallel at the same time with the training of a second individual sub-task of the decomposed task incorporated into its own hierarchical graph of concept nodes.