公开(公告)号：US12240117B2

公开(公告)日：2025-03-04

申请号：US18157919

申请日：2023-01-23

Applicant: Google LLC

Inventor： Yevgen Chebotar ,  Pierre Sermanet ,  Harrison Lynch

IPC: B25J9/16 ,  G05B13/02 ,  G06N3/084 ,  G06N20/00

Abstract: There are provided systems, methods, and apparatus, for optimizing a policy controller to control a robotic agent that interacts with an environment to perform a robotic task. One of the methods includes optimizing the policy controller using a neural network that generates numeric embeddings of images of the environment and a demonstration sequence of demonstration images of another agent performing a version of the robotic task.

公开(公告)号：US11559887B2

公开(公告)日：2023-01-24

申请号：US16649596

申请日：2018-09-20

Applicant: GOOGLE LLC

Inventor： Yevgen Chebotar ,  Pierre Sermanet ,  Harrison Lynch

IPC: G06N20/00 ,  G05B13/02 ,  G06N3/08 ,  B25J9/16

Abstract: There are provided systems, methods, and apparatus, for optimizing a policy controller to control a robotic agent that interacts with an environment to perform a robotic task. One of the methods includes optimizing the policy controller using a neural network that generates numeric embeddings of images of the environment and a demonstration sequence of demonstration images of another agent performing a version of the robotic task.

公开(公告)号：US20200276703A1

公开(公告)日：2020-09-03

申请号：US16649596

申请日：2018-09-20

Applicant: GOOGLE LLC

Inventor： Yevgen Chebotar ,  Pierre Sermanet ,  Harrison Lynch

IPC: B25J9/16 ,  G05B13/02 ,  G06N3/08 ,  G06N20/00

Abstract: There are provided systems, methods, and apparatus, for optimizing a policy controller to control a robotic agent that interacts with an environment to perform a robotic task. One of the methods includes optimizing the policy controller using a neural network that generates numeric embeddings of images of the environment and a demonstration sequence of demonstration images of another agent performing a version of the robotic task.