Deepmind technologies limited (20250124297). CONTROLLING REINFORCEMENT LEARNING AGENTS USING GEOMETRIC POLICY COMPOSITION
CONTROLLING REINFORCEMENT LEARNING AGENTS USING GEOMETRIC POLICY COMPOSITION
Organization Name
Inventor(s)
Mark Daniel Rowland of London GB
Shantanu Yogeshraj Thakoor of London GB
Andre Da Motta Salles Barreto of London GB
Diana Luiza Borsa of London GB
William Clinton Dabney of London GB
CONTROLLING REINFORCEMENT LEARNING AGENTS USING GEOMETRIC POLICY COMPOSITION
This abstract first appeared for US patent application 20250124297 titled 'CONTROLLING REINFORCEMENT LEARNING AGENTS USING GEOMETRIC POLICY COMPOSITION
Original Abstract Submitted
methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for controlling a reinforcement learning agent in an environment. one of the methods may include maintaining data specifying a base policy set comprising a plurality of base policies for controlling the agent; receiving a current observation characterizing a current state of the environment; generating, for each of the plurality of base policies, one or more predicted future observations characterizing respective future states of the environment that are subsequent to the current state of the environment; using the predicted future observations generated for the plurality of base policies to determine a respective estimated value for each composite policy in a composite policy set with respect to the current state of the environment; and selecting an action using the respective estimated values for the composite policies.
