US 12,321,719 B1
System and method for preprocessing and conditioning cosmic ray data for randomness generation
Scott Streit, Cheyenne, WY (US); David Harding, Cheyenne, WY (US); Matt Klepp, Cheyenne, WY (US); and Richard Kane, Cheyenne, WY (US)
Assigned to Entrokey Labs Inc., Cheyenne, WY (US)
Filed by ENTROKEY LABS INC., Cheyenne, WY (US)
Filed on Dec. 20, 2024, as Appl. No. 18/991,130.
Int. Cl. G06F 7/58 (2006.01); G06N 3/045 (2023.01); G06N 3/0475 (2023.01); G06N 3/094 (2023.01)
CPC G06F 7/588 (2013.01) [G06N 3/045 (2023.01); G06N 3/0475 (2023.01); G06N 3/094 (2023.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for preprocessing and conditioning raw cosmic ray data to generate a binary sequence representing randomness, the method comprising:
receiving a plurality of electrical signals generated in response to detecting raw cosmic ray data captured during at least one cosmic ray detection event;
extracting a set of values associated with each electrical signal, each set comprising at least one of a pulse amplitude value, a signal amplitude value, and a timestamp value;
preprocessing each value within the extracted set;
conditioning each of the preprocessed values to generate a random binary sequence; and
feeding the conditioned random binary sequence to train a generator of a neural network to generate a random sequence mathematically similar to the conditioned random binary sequence; and
feeding the generated random sequence to train a discriminator of the neural network to distinguish the generated random binary sequence from the conditioned random binary sequence;
wherein the discriminator assigns a score to the generated random sequence characterizing randomness of the generated random binary sequence.