US 11,967,141 B2
Neural architecture search for fusing multiple networks into one
Adrien David Gaidon, Mountain View, CA (US); and Jie Li, Los Altos, CA (US)
Assigned to TOYOTA RESEARCH INSTITUTE, INC., Los Altos, CA (US)
Filed by TOYOTA RESEARCH INSTITUTE, INC., Los Altos, CA (US)
Filed on Jan. 30, 2023, as Appl. No. 18/161,777.
Application 18/161,777 is a continuation of application No. 16/853,181, filed on Apr. 20, 2020, granted, now 11,568,210.
Prior Publication US 2023/0177825 A1, Jun. 8, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 3/04 (2023.01); G06F 17/18 (2006.01); G06F 18/20 (2023.01); G06F 18/2113 (2023.01); G06F 18/25 (2023.01); G06N 3/045 (2023.01); G06V 10/764 (2022.01); G06V 10/771 (2022.01); G06V 10/80 (2022.01); G06V 10/82 (2022.01); G06V 20/56 (2022.01); G06V 40/10 (2022.01)
CPC G06V 10/82 (2022.01) [G06F 17/18 (2013.01); G06F 18/2113 (2023.01); G06F 18/25 (2023.01); G06F 18/29 (2023.01); G06N 3/045 (2023.01); G06V 10/764 (2022.01); G06V 10/771 (2022.01); G06V 10/80 (2022.01); G06V 20/56 (2022.01); G06V 40/10 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A vehicle, comprising:
a computer system, wherein the computer system:
generates a directed acyclic graph representing at least a partial union of respective architectures associated with multiple trained networks, each of the multiple trained networks being associated with a respective task and the directed acyclic graph being associated with multiple tasks relating to the multiple trained networks;
defines a joint objective for the direct acyclic graph; and
converges the joint objective for the multiple tasks over the direct acyclic graph; and
a controller device, wherein the controller device:
receives the directed acyclic graph from the computer system; and
performs autonomous operations controlling the vehicle based on machine learning associated with the directed acyclic graph.