US 12,067,114 B2
Byte n-gram embedding model
Radu Cazan, Iasi (RO); Daniel Radu, Bucharest (RO); and Marian Radu, Bucharest (RO)
Assigned to CrowdStrike, Inc., Sunnyvale, CA (US)
Filed by CrowdStrike, Inc., Irvine, CA (US)
Filed on Jun. 22, 2023, as Appl. No. 18/213,141.
Application 18/213,141 is a continuation of application No. 16/237,468, filed on Dec. 31, 2018, granted, now 11,727,112.
Claims priority of provisional application 62/692,331, filed on Jun. 29, 2018.
Prior Publication US 2023/0334154 A1, Oct. 19, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 21/56 (2013.01); G06F 18/214 (2023.01); G06F 21/55 (2013.01); G06N 3/08 (2023.01)
CPC G06F 21/56 (2013.01) [G06F 18/214 (2023.01); G06F 21/552 (2013.01); G06N 3/08 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving data to be processed;
identifying, from the data, a plurality of byte n-grams using a pseudo-random number generator;
generating, using a neural network, a hash of the data based at least in part on the plurality of byte n-grams;
classifying the data into a predefined classification based at least in part on the hash; and
generating a signature of the data based at least in part on the hash and information associated with the predefined classification.