US 11,714,877 B1
System for training of recognition system using ad hoc training data
Alon Shoshan, Haifa (IL); Miriam Farber, Haifa (IL); Nadav Israel Bhonker, Haifa (IL); Igor Kviatkovsky, Haifa (IL); Manoj Aggarwal, Seattle, WA (US); and Gerard Guy Medioni, Seattle, WA (US)
Assigned to AMAZON TECHNOLOGIES, INC., Seattle, WA (US)
Filed by AMAZON TECHNOLOGIES, INC., Seattle, WA (US)
Filed on Sep. 30, 2020, as Appl. No. 17/38,463.
Int. Cl. G06K 9/62 (2022.01); G06N 3/08 (2006.01); G06N 3/04 (2006.01); G06V 10/145 (2022.01); G06V 40/13 (2022.01); G06N 3/088 (2023.01); G06V 40/12 (2022.01); G06N 3/045 (2023.01); G06F 18/214 (2023.01)
CPC G06F 18/214 (2023.01) [G06N 3/0454 (2013.01); G06N 3/088 (2013.01); G06V 10/145 (2022.01); G06V 40/1318 (2022.01); G06V 40/1347 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a memory, storing first computer-executable instructions; and
one or more hardware processors to execute the first computer-executable instructions to:
acquire actual data comprising a first set of images;
determine a first plurality of triplets, wherein each triplet of the first plurality of triplets comprises at least a portion of the actual data;
generate a second plurality of triplets, wherein at least one triplet of the second plurality of triplets comprises first synthetic data;
train, at a first time and using the first plurality of triplets and the second plurality of triplets, a machine learning system that uses a first loss function, wherein the machine learning system generates recognition data that is indicative of determines an identity of a user;
determine first assessment data indicative of performance of the machine learning system trained at the first time;
based on the first assessment data, generate a first training request third plurality of triplets, wherein each triplet of the third plurality of triplets comprises second synthetic data; and
train, at a second time and using the first plurality of triplets and the third plurality of triplets, the machine learning system.