Deepmind Technologies Limited

Gdm Holding LLC

Nicolas Manfred Otto Heess

Oriol Vinyals

Koray Kavukcuoglu

Timothy Paul Lillicrap

Mel Vecerik

Raia Thais Hadsell

Daniel Pieter Wierstra

Yuval Tassa

Gregory Duncan Wayne

Yujia Li

Gabriel Barth-Maron

David Silver

Joao Ferdinando Gomes De Freitas

Razvan Pascanu

Volodymyr Mnih

Martin Riedmiller

Thomas Lampe

Peter William Battaglia

Victor Constant Bapst

Thomas Rothoerl

Theophane Guillaume Weber

Sebastien Henri Andre Racaniere

Roland Hafner

Lars Buesing

Wojciech Czarnecki

Danilo Jimenez Rezende

Adrià Puigdomènech Badia

Alexander Pritzel

Ziyu Wang

James Kirkpatrick

Yuke Zhu

Remi Munos

Andrei-Alexandru Rusu

Todd Andrew Hester

Ivaylo Popov

David Paul Reichert

Jonathan Karl Scholz

Bilal Piot

Olivier Claude Pietquin

Yee Whye Teh

Arthur Clement Guez

Tom Erez

Jonathan James Hunt

Wang Fumin

Joshua Merel

Alexandre Galashov

Adria Puigdomenech Badia

Rae Chan Jeong

Alvaro Sanchez

Saran Tunyasuvunakool

János Kramár

Francesco Nori

Jost Tobias Springenberg

Victore Bapst

Jacqueline Ok-Chan Kay

Danilo Jiminez Rezende

Matej Vecerik

Olivier Pietquin

Scott Ellison Reed

Joao Ferdinando Freitas

Sebastien Henri Racaniere

A graph neural network system for processing data representing a physical system is disclosed. The graph neural network system receives physical system data and action data, where the physical system includes bodies coupled by joints. The graph neural network system processes an input graph to provide an output graph, the input and output graphs each having a graph structure including nodes and edges corresponding, respectively, to the bodies and joints of the physical system. The graph processing neural network provides the inferred static property or dynamic state of the physical system. The graph processing neural network is differentiable and may be used for control and/or reinforcement learning.

Graph neural networks representing physical systems

Methods, systems, and apparatus for selecting actions to be performed by an agent interacting with an environment. One system includes a high-level controller neural network, low-level controller network, and subsystem. The high-level controller neural network receives an input observation and processes the input observation to generate a high-level output defining a control signal for the low-level controller. The low-level controller neural network receives a designated component of an input observation and processes the designated component and an input control signal to generate a low-level output that defines an action to be performed by the agent in response to the input observation. The subsystem receives a current observation characterizing a current state of the environment, determines whether criteria are satisfied for generating a new control signal, and based on the determination, provides appropriate inputs to the high-level and low-level controllers for selecting an action to be performed by the agent.

Selecting reinforcement learning actions using a low-level controller

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for data-efficient reinforcement learning. One of the systems is a system for training an actor neural network used to select actions to be performed by an agent that interacts with an environment by receiving observations characterizing states of the environment and, in response to each observation, performing an action selected from a continuous space of possible actions, wherein the actor neural network maps observations to next actions in accordance with values of parameters of the actor neural network, and wherein the system comprises: a plurality of workers, wherein each worker is configured to operate independently of each other worker, wherein each worker is associated with a respective agent replica that interacts with a respective replica of the environment during the training of the actor neural network.

Data-efficient reinforcement learning for continuous control tasks

A neural network control system for controlling an agent to perform a task in a real-world environment, operates based on both image data and proprioceptive data describing the configuration of the agent. The training of the control system includes both imitation learning, using datasets generated from previous performances of the task, and reinforcement learning, based on rewards calculated from control data output by the control system.

Reinforcement and imitation learning for a task

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an action selection neural network. One of the methods includes maintaining a replay memory that stores trajectories generated as a result of interaction of an agent with an environment; and training an action selection neural network having policy parameters on the trajectories in the replay memory, wherein training the action selection neural network comprises: sampling a trajectory from the replay memory; and adjusting current values of the policy parameters by training the action selection neural network on the trajectory using an off-policy actor critic reinforcement learning technique.

Training action selection neural networks using off-policy actor critic reinforcement learning and stochastic dueling neural networks

A method is proposed for training a multitask computer system, such as a multitask neural network system. The system comprises a set of trainable workers and a shared module. The trainable workers and shared module are trained on a plurality of different tasks, such that each worker learns to perform a corresponding one of the tasks according to a respective task policy, and said shared policy network learns a multitask policy which represents common behavior for the tasks. The coordinated training is performed by optimizing an objective function comprising, for each task: a reward term indicative of an expected reward earned by a worker in performing the corresponding task according to the task policy; and at least one entropy term which regularizes the distribution of the task policy towards the distribution of the multitask policy.

Multi-task neural network systems with task-specific policies and a shared policy

An off-policy reinforcement learning actor-critic neural network system configured to select actions from a continuous action space to be performed by an agent interacting with an environment to perform a task. An observation defines environment state data and reward data. The system has an actor neural network which learns a policy function mapping the state data to action data. A critic neural network learns an action-value (Q) function. A replay buffer stores tuples of the state data, the action data, the reward data and new state data. The replay buffer also includes demonstration transition data comprising a set of the tuples from a demonstration of the task within the environment. The neural network system is configured to train the actor neural network and the critic neural network off-policy using stored tuples from the replay buffer comprising tuples both from operation of the system and from the demonstration transition data.

Training action selection neural networks using apprenticeship

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training an actor neural network used to select actions to be performed by an agent interacting with an environment. One of the methods includes obtaining a minibatch of experience tuples; and updating current values of the parameters of the actor neural network, comprising: for each experience tuple in the minibatch: processing the training observation and the training action in the experience tuple using a critic neural network to determine a neural network output for the experience tuple, and determining a target neural network output for the experience tuple; updating current values of the parameters of the critic neural network using errors between the target neural network outputs and the neural network outputs; and updating the current values of the parameters of the actor neural network using the critic neural network.

Continuous control with deep reinforcement learning

A system includes a neural network system implemented by one or more computers. The neural network system is configured to receive an observation characterizing a current state of a real-world environment being interacted with by a robotic agent to perform a robotic task and to process the observation to generate a policy output that defines an action to be performed by the robotic agent in response to the observation. The neural network system includes: (i) a sequence of deep neural networks (DNNs), in which the sequence of DNNs includes a simulation-trained DNN that has been trained on interactions of a simulated version of the robotic agent with a simulated version of the real-world environment to perform a simulated version of the robotic task, and (ii) a first robot-trained DNN that is configured to receive the observation and to process the observation to generate the policy output.

Neural networks for selecting actions to be performed by a robotic agent

A neural network system is proposed. The neural network can be trained by model-based reinforcement learning to select actions to be performed by an agent interacting with an environment, to perform a task in an attempt to achieve a specified result. The system may comprise at least one imagination core which receives a current observation characterizing a current state of the environment, and optionally historical observations, and which includes a model of the environment. The imagination core may be configured to output trajectory data in response to the current observation, and/or historical observations. The trajectory data comprising a sequence of future features of the environment imagined by the imagination core. The system may also include a rollout encoder to encode the features, and an output stage to receive data derived from the rollout embedding and to output action policy data for identifying an action based on the current observation.

Imagination-based agent neural networks

Growing community of inventors

London, United Kingdom

Nicolas Manfred Otto Heess