US 12,347,180 B1
System and method for pattern recognition and graph extraction for data collected from gate-defined semiconductor quantum dots
Joseph Comer, Portland, OR (US); Reed Andrews, Port Hueneme, CA (US); Ian Jenkins, Westlake Village, CA (US); Parker Williams, Calabasas, CA (US); Heiko Hoffmann, Simi Valley, CA (US); and Wyatt Mcallister, Malibu, CA (US)
Assigned to HRL LABORATORIES, LLC, Malibu, CA (US)
Filed by HRL Laboratories, LLC, Malibu, CA (US)
Filed on Feb. 28, 2023, as Appl. No. 18/115,549.
Claims priority of provisional application 63/315,907, filed on Mar. 2, 2022.
Int. Cl. G06V 10/84 (2022.01); G06V 10/75 (2022.01); G06V 10/82 (2022.01)
CPC G06V 10/84 (2022.01) [G06V 10/751 (2022.01); G06V 10/82 (2022.01)] 6 Claims
OG exemplary drawing
 
1. A system for pattern recognition and graph extraction, the system comprising:
one or more processors and associated memory, the memory being a non-transitory computer-readable medium having executable instructions encoded thereon, such that upon execution of the instructions, the one or more processors perform operations of:
obtaining data from a quantum dot device having gate defined semiconductor quantum dots;
generating an image with pixel-level coordinates from the data;
receiving graph annotations of the image;
processing, with a deep network, the image and pixel-level coordinates to generate a predicted graph;
comparing the predicted graph to the graph annotations to generate a loss;
optimizing the deep network by updating parameters in the deep network based on the loss;
processing, with the optimized deep network, the image with pixel-level coordinates to generate an optimized predicted graph;
identifying, with the optimized predicted graph, operational voltages to apply to the quantum dot device; and
applying the operational voltages to the quantum dot device.