Microsoft Technology Licensing, LLC

Sam Michael Devlin

Katja Hofmann

Kamil Andrzej Ciosek

Sebastian Tschiatschek

Cheng Zhang

Yingzhen Li

Maximilian Igl

Adrian Kieron O'Grady

Anthony David Joseph Diggle

Andrew Philip Slowey

Janhavi Agrawal

Luisa Maria Zintgraf

Jay Nanavati

David Michael Bignell

An apparatus has a memory storing a reinforcement learning policy with an optimization component and a data collection component. The apparatus has a regularization component which applies regularization selectively between the optimization component of the reinforcement learning policy and the data collection component of the reinforcement learning policy. A processor carries out a reinforcement learning process by: triggering execution of an agent according to the policy and with respect to a first task; observing values of variables comprising: an observation space of the agent, an action of the agent; and updating the policy using reinforcement learning according to the observed values and taking into account the regularization.

Behavior policy for a generalized reinforcement learning agent utilizing a weight of a loss function

In various examples there is a method performed by a computer-implemented agent in an environment. The method comprises storing a reinforcement learning policy for controlling the computer-implemented agent. The method also comprises storing a distribution as a latent representation of a belief of the computer-implemented agent about at least one other agent in the environment. The method involves executing the computer-implemented agent according to the policy conditioned on parameters characterizing the distribution.

Interactive agent and control using reinforcement learning

Generalized reinforcement learning agent

An apparatus is described for training a behavior of an agent in a physical or digital environment. The apparatus comprises a memory storing the location of at least one reward token in the environment. The location has been specified by a user. At least one processor executes the agent in the environment according to a behavior policy. The processor is configured to observe values of variables comprising: an observation of the agent, an action of the agent and any reward resulting from the reward token. The processor is configured to update the behavior policy using reinforcement learning according to the observed values.

Training behavior of an agent

Growing community of inventors

Trumpington, United Kingdom

Sam Michael Devlin