公开(公告)号：US11907821B2

公开(公告)日：2024-02-20

申请号：US16586236

申请日：2019-09-27

Applicant: DeepMind Technologies Limited

Inventor： Ang Li ,  Valentin Clement Dalibard ,  David Budden ,  Ola Spyra ,  Maxwell Elliot Jaderberg ,  Timothy James Alexander Harley ,  Sagi Perel ,  Chenjie Gu ,  Pramod Gupta

IPC: G06N20/20 ,  G06F16/901 ,  G06N5/04

CPC classification number: G06N20/20 ,  G06F16/9024 ,  G06N5/04

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a machine learning model. A method includes: maintaining a plurality of training sessions; assigning, to each worker of one or more workers, a respective training session of the plurality of training sessions; repeatedly performing operations until meeting one or more termination criteria, the operations comprising: receiving an updated training session from a respective worker of the one or more workers, selecting a second training session, selecting, based on comparing the updated training session and the second training session using a fitness evaluation function, either the updated training session or the second training session as a parent training session, generating a child training session from the selected parent training session, and assigning the child training session to an available worker, and selecting a candidate model to be a trained model for the machine learning model.

公开(公告)号：US20230281445A1

公开(公告)日：2023-09-07

申请号：US18120715

申请日：2023-03-13

Applicant: DeepMind Technologies Limited

Inventor： Maxwell Elliot Jaderberg ,  Wojciech Czarnecki ,  Timothy Frederick Goldie Green ,  Valentin Clement Dalibard

IPC: G06N3/08 ,  G06N3/048

CPC classification number: G06N3/08 ,  G06N3/048

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. A method includes: training a neural network having a plurality of network parameters to perform a particular neural network task and to determine trained values of the network parameters using an iterative training process having a plurality of hyperparameters, the method comprising: maintaining a plurality of candidate neural networks and, for each of the candidate neural networks, data specifying: (i) respective values of the network parameters for the candidate neural network, (ii) respective values of the hyperparameters for the candidate neural network, and (iii) a quality measure that measures a performance of the candidate neural network on the particular neural network task; and for each of the plurality of candidate neural networks, repeatedly performing additional training operations.

公开(公告)号：US20210004676A1

公开(公告)日：2021-01-07

申请号：US16766631

申请日：2018-11-22

Applicant: DeepMind Technologies Limited

Inventor： Maxwell Elliot Jaderberg ,  Wojciech Czarnecki ,  Timothy Frederick Goldie Green ,  Valentin Clement Dalibard

IPC: G06N3/08 ,  G06N3/04

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. A method includes: training a neural network having a plurality of network parameters to perform a particular neural network task and to determine trained values of the network parameters using an iterative training process having a plurality of hyperparameters, the method comprising: maintaining a plurality of candidate neural networks and, for each of the candidate neural networks, data specifying: (i) respective values of the network parameters for the candidate neural network, (ii) respective values of the hyperparameters for the candidate neural network, and (iii) a quality measure that measures a performance of the candidate neural network on the particular neural network task; and for each of the plurality of candidate neural networks, repeatedly performing additional training operations.

公开(公告)号：US20240242091A1

公开(公告)日：2024-07-18

申请号：US18562180

申请日：2022-05-30

Applicant: DeepMind Technologies Limited

Inventor： Valentin Clement Dalibard ,  Maxwell Elliot Jaderberg

IPC: G06N3/0985 ,  G06N3/045

CPC classification number: G06N3/0985 ,  G06N3/045

Abstract: Methods, computer systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network for performing a task. The system maintains data specifying (i) a plurality of candidate neural networks and (ii) a partitioning of the plurality of candidate neural networks into a plurality of partitions. The system repeatedly performs operations, including: training each of the candidate neural networks; evaluating each candidate neural network using a respective fitness function for the partition; and for each partition, updating the respective values of the one or more hyperparameters for at least one of the candidate neural networks in the partition based on the respective fitness metrics of the candidate neural networks in the partition. After repeatedly performing the operations, the system selects, from the maintained data, the respective values of the network parameters of one of the candidate neural networks.

公开(公告)号：US20210097443A1

公开(公告)日：2021-04-01

申请号：US16586236

申请日：2019-09-27

Applicant: DeepMind Technologies Limited

Inventor： Ang Li ,  Valentin Clement Dalibard ,  David Budden ,  Ola Spyra ,  Maxwell Elliot Jaderberg ,  Timothy James Alexander Harley ,  Sagi Perel ,  Chenjie Gu ,  Pramod Gupta

IPC: G06N20/20 ,  G06N5/04 ,  G06F16/901

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a machine learning model. A method includes: maintaining a plurality of training sessions; assigning, to each worker of one or more workers, a respective training session of the plurality of training sessions; repeatedly performing operations until meeting one or more termination criteria, the operations comprising: receiving an updated training session from a respective worker of the one or more workers, selecting a second training session, selecting, based on comparing the updated training session and the second training session using a fitness evaluation function, either the updated training session or the second training session as a parent training session, generating a child training session from the selected parent training session, and assigning the child training session to an available worker, and selecting a candidate model to be a trained model for the machine learning model.

公开(公告)号：US20240346310A1

公开(公告)日：2024-10-17

申请号：US18612917

申请日：2024-03-21

Applicant: DeepMind Technologies Limited

Inventor： Maxwell Elliot Jaderberg ,  Wojciech Czarnecki ,  Timothy Frederick Goldie Green ,  Valentin Clement Dalibard

IPC: G06N3/08 ,  G06N3/048

CPC classification number: G06N3/08 ,  G06N3/048

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. A method includes: training a neural network having a plurality of network parameters to perform a particular neural network task and to determine trained values of the network parameters using an iterative training process having a plurality of hyperparameters, the method comprising: maintaining a plurality of candidate neural networks and, for each of the candidate neural networks, data specifying: (i) respective values of the network parameters for the candidate neural network, (ii) respective values of the hyperparameters for the candidate neural network, and (iii) a quality measure that measures a performance of the candidate neural network on the particular neural network task; and for each of the plurality of candidate neural networks, repeatedly performing additional training operations.

公开(公告)号：US20240127071A1

公开(公告)日：2024-04-18

申请号：US18475859

申请日：2023-09-27

Applicant: DeepMind Technologies Limited

Inventor： Robert Tjarko Lange ,  Tom Schaul ,  Yutian Chen ,  Tom Ben Zion Zahavy ,  Valentin Clement Dalibard ,  Christopher Yenchuan Lu ,  Satinder Singh Baveja ,  Johan Sebastian Flennerhag

IPC: G06N3/086

CPC classification number: G06N3/086

Abstract: There is provided a computer-implemented method for updating a search distribution of an evolutionary strategies optimizer using an optimizer neural network comprising one or more attention blocks. The method comprises receiving a plurality of candidate solutions, one or more parameters defining the search distribution that the plurality of candidate solutions are sampled from, and fitness score data indicating a fitness of each respective candidate solution of the plurality of candidate solutions. The method further comprises processing, by the one or more attention neural network blocks, the fitness score data using an attention mechanism to generate respective recombination weights corresponding to each respective candidate solution. The method further comprises updating the one or more parameters defining the search distribution based upon the recombination weights applied to the plurality of candidate solutions.

公开(公告)号：US11941527B2

公开(公告)日：2024-03-26

申请号：US18120715

申请日：2023-03-13

Applicant: DeepMind Technologies Limited

Inventor： Maxwell Elliot Jaderberg ,  Wojciech Czarnecki ,  Timothy Frederick Goldie Green ,  Valentin Clement Dalibard

IPC: G06N3/08 ,  G06N3/048

CPC classification number: G06N3/08 ,  G06N3/048

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. A method includes: training a neural network having a plurality of network parameters to perform a particular neural network task and to determine trained values of the network parameters using an iterative training process having a plurality of hyperparameters, the method comprising: maintaining a plurality of candidate neural networks and, for each of the candidate neural networks, data specifying: (i) respective values of the network parameters for the candidate neural network, (ii) respective values of the hyperparameters for the candidate neural network, and (iii) a quality measure that measures a performance of the candidate neural network on the particular neural network task; and for each of the plurality of candidate neural networks, repeatedly performing additional training operations.

公开(公告)号：US11604985B2

公开(公告)日：2023-03-14

申请号：US16766631

申请日：2018-11-22

Applicant: DeepMind Technologies Limited

Inventor： Maxwell Elliot Jaderberg ,  Wojciech Czarnecki ,  Timothy Frederick Goldie Green ,  Valentin Clement Dalibard

IPC: G06N3/08 ,  G06N3/04

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. A method includes: training a neural network having multiple network parameters to perform a particular neural network task and to determine trained values of the network parameters using an iterative training process having multiple hyperparameters, the method includes: maintaining multiple candidate neural networks and, for each of the multiple candidate neural networks, data specifying: (i) respective values of network parameters for the candidate neural network, (ii) respective values of hyperparameters for the candidate neural network, and (iii) a quality measure that measures a performance of the candidate neural network on the particular neural network task; and for each of the multiple candidate neural networks, repeatedly performing additional training operations.