US 12,443,881 B1
Apparatus and methods to provide a learning agent with improved computational applications in complex real-world environments using machine learning
James Brennan Worth, Castellón de la Plana (ES)
Assigned to Substrate Artificial Intelligence SA, Madrid (ES)
Filed by Substrate Artificial Intelligence SA, Madrid (ES)
Filed on Sep. 2, 2022, as Appl. No. 17/902,455.
Claims priority of provisional application 63/240,022, filed on Sep. 2, 2021.
Int. Cl. G06F 16/00 (2019.01); G06N 20/00 (2019.01)
CPC G06N 20/00 (2019.01) 22 Claims
OG exemplary drawing
 
1. A method, comprising:
building a machine learning (ML) model that includes a digital agent configured to navigate a digital environment;
implementing a navigation of the digital agent over a period of time and following a path within the digital environment, 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 upon reaching each state configured to acquire the value associated with that state;
determining a signal representing an acquisition of values by the digital agent over the period of time;
computing, based on the signal representing the acquisition of values, an emotive state of the digital agent;
computing, based on the signal representing the acquisition of values, an arousal state of the digital agent;
determining a goal for the digital agent based on at least one of the emotive state or the arousal state of the digital agent, the goal configured such that the digital agent achieves an improved overall performance in acquiring cumulative rewards when the digital agent takes actions to navigate towards the goal compared to when the digital agent does not take actions to navigate towards the goal; and
receive an indication of action taken by the digital agent in the digital environment and implement a real version of the action in a real environment, the digital environment being associated with the real environment.