Apparatus and methods to provide a learning agent with improved computational applications in complex real-world environments using machine learning

An apparatus, comprising:

• a memory;
• and
• a hardware processor operatively coupled to the memory, the hardware processor configured to:
• implement, using a stack of ML models according to a hierarchical ML model architecture, a navigation of a digital agent in a digital environment towards a predefined goal over a period of time and following a path, the path including a set of states, each state from the set of states being associated with a value from a set of values, the digital agent configured to, upon reaching each state, acquire the value associated with that state and from the set of values;
• monitor a signal representing an acquisition of values from the set of values over the period of time and by the digital agent;
• calculate a sub-goal based on a rate of acquisition of values and using a first ML model from the stack of ML models, the sub-goal configured to adjust the digital agent's behavior to finding a balance between exploration for unknown values and exploitation of known values, the first ML model being at a first hierarchical level in the hierarchical ML model architecture;
• and
• generating an agent action, using a second ML model from the stack of ML models different than the first ML model and at a second hierarchical level in the hierarchical ML model architecture different than the first hierarchical level, and based on the sub-goal, the sub-goal being configured such that the digital agent achieves an improved overall performance in acquiring cumulative rewards when the digital agent performs actions to navigate towards the sub-goal compared to when the digital agent does not perform actions to navigate towards the sub-goal.

A non-transitory processor-readable medium storing code representing instructions to be executed by a processor, the instructions comprising code to cause the processor to:

• receive a signal representing an acquisition of values by a digital agent navigating a digital environment, the digital agent being directed to navigate towards a predefined goal;
• calculate an emotive state of the digital agent based on the signal representing the acquisition of values by the digital agent;
• calculate an arousal state of the digital agent based on the signal representing the acquisition of values by the digital agent;
• generate a sub-goal, based on the emotive state and the arousal state of the digital agent, the sub-goal configured to at least one of direct the digital agent toward the predefined goal or increase a cumulative reward acquired by the digital agent;
• receive an indication of action performed by the digital agent in the digital environment, based on the sub-goal;
• and
• implement a real version of the action in a real environment, the digital environment being associated with the real environment.