US 12,243,226 B2
Bi-directional quantum annealing in markov random fields for machine learning in image analysis
Shreyas Ramesh, Mountainside, NJ (US); Kung-Chuan Hsu, Cerritos, CA (US); Max Howard, San Francisco, CA (US); and Hassan Naseri, Helsinki (FI)
Assigned to Accenture Global Solutions Limited, Dublin (IE)
Filed by Accenture Global Solutions Limited, Dublin (IE)
Filed on Mar. 9, 2022, as Appl. No. 17/690,662.
Claims priority of provisional application 63/159,786, filed on Mar. 11, 2021.
Prior Publication US 2022/0292675 A1, Sep. 15, 2022
Int. Cl. G06T 7/00 (2017.01); G06N 3/044 (2023.01); G06N 3/047 (2023.01); G06N 3/08 (2023.01); G06N 10/20 (2022.01); G06V 10/764 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)
CPC G06T 7/0012 (2013.01) [G06N 3/044 (2023.01); G06N 3/047 (2023.01); G06N 3/08 (2013.01); G06N 10/20 (2022.01); G06V 10/764 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)] 21 Claims
OG exemplary drawing
 
1. A computer implemented method for training an image classifier, comprising:
obtaining training data comprising features extracted from a first set of images;
training a deep quantum restricted Boltzmann machine (QRBM) comprising multiple layers using the training data, the training comprising layer-wise training of the multiple layers, wherein each layer of the multiple layers included in the deep QRBM comprises a restricted Boltzmann machine (RBM) and training each layer of the multiple layers comprises evaluating the RBM probability distribution using bi-directional quantum annealing;
validating the trained deep QRBM using test data comprising features extracted from a second set of images; and
evaluating, using the bi-directional quantum annealing, probability distributions in Markov Random Field (MRF) models,
wherein a quantum computing resource implements the bi-directional annealing.