Machine learning identification of quantum vulnerable cryptography
Systems, devices, methods, and computer-readable media provide automatic detection of cryptographic algorithms susceptible to compromise by a quantum computer. A method includes receiving, by a trained machine learning (ML) model, a file; executing the trained ML model on the file to generate an output, the output indicating, for each cryptography algorithm of a plurality of cryptography algorithms that are susceptible to compromise by a quantum computer, whether the cryptography algorithm is embodied in or used by the file. Responsive to determining at least one of the cryptography algorithms is embodied or used in the file, performing a mitigation action that reduces or eliminates a vulnerability of the file to compromise by a quantum computer.
1 . A method comprising:
receiving, by a trained machine learning (ML) model, a file comprising contents including source code, a function call, or mathematical operations;
executing the trained ML model on the contents of the file to generate an output comprising a classification vector, the output indicating, for each cryptography algorithm of a plurality of asymmetric cryptography algorithms that are susceptible to compromise by a quantum computer, whether the cryptography algorithm is embodied in or used by the file; and
responsive to determining, based on the classification vector, that at least one of the cryptography algorithms is embodied or used in the file, replacing the contents in the file associated with the cryptography algorithm with corresponding contents that are not susceptible to compromise by a quantum computer.
2 . The method of claim 1 , wherein the ML model comprises a deep neural network (DNN).
3 . The method of claim 1 , wherein the ML model comprises a support vector machine (SVM) model.
4 . The method of claim 1 , further comprising, by executing the trained ML model on the contents of the file, identifying features of each of the plurality of cryptography algorithms;
the features, when present in the file, indicating the presence of quantum vulnerable cryptography algorithms.
5 . The method of claim 4 , wherein the features include text representative of a name of the cryptography algorithm, a number used in the cryptography algorithm, and/or text representative of a function called by the cryptography algorithm.
6 . The method of claim 1 , further comprising training the ML model based on files associated with a label vector comprising, for each asymmetric cryptography algorithm susceptible to compromise by a quantum computer, a present/not-present indication.
7 . The method of claim 1 , wherein the trained ML model comprises a plurality of ML models, each of the ML models corresponding to the plurality of cryptography algorithms susceptible to compromise by a quantum computer, wherein each ML model is a binary classifier trained to output a value for a respective asymmetric cryptography algorithm indicating whether that algorithm is embodied in or used by the file.
8 . A non-transitory machine-readable medium including instructions that, when executed by a machine, cause the machine to perform operations for identifying a cryptography algorithm, in a file, that is vulnerable to compromise by a quantum computer, the operations comprising:
receiving, by a trained machine learning (ML) model, a file comprising contents including source code, a function call, or mathematical operations; and
executing the trained ML model on the contents of the file to generate an output comprising a classification vector, the output indicating, for each cryptography algorithm of a plurality of asymmetric cryptography algorithms that are susceptible to compromise by the quantum computer, whether the cryptography algorithm is embodied in or used by the file; and
responsive to determining, based on the classification vector, that at least one of the cryptography algorithms is embodied or used in the file, replacing the contents in the file associated with the cryptography algorithm with corresponding contents that are not susceptible to compromise by a quantum computer.
9 . The non-transitory machine-readable medium of claim 8 , wherein the ML model is a deep neural network (DNN).
10 . The non-transitory machine-readable medium of claim 8 , wherein the ML model is a support vector machine (SVM) model.
11 . The non-transitory machine-readable medium of claim 8 , further comprising, by executing the trained ML model on the contents of the file, identifying features of each of the plurality of cryptography algorithms;
the features, when present in the file, indicating the presence of quantum vulnerable cryptography algorithms.
12 . The non-transitory machine-readable medium of claim 11 , wherein the features include text representative of a name of the cryptography algorithm, a number used in the cryptography algorithm, and/or text representative of a function called by the cryptography algorithm.
13 . The non-transitory machine-readable medium of claim 8 , further comprising training the ML model based on files associated with a label vector comprising, for each asymmetric cryptography algorithm susceptible to compromise by a quantum computer, a present/not-present indication.
14 . A system comprising:
processing circuitry; and
a memory including instructions that, when executed by the processing circuitry, cause the processing circuitry to perform operations comprising:
receiving, by a trained machine learning (ML) model, a file comprising contents including source code, a function call, or mathematical operations; and
executing the trained ML model on the contents of the file to generate an output comprising a classification vector, the output indicating, for each cryptography algorithm of a plurality of asymmetric cryptography algorithms that are susceptible to compromise by a quantum computer, whether the cryptography algorithm is embodied in or used by the file; and
responsive to determining, based on the classification vector, that at least one of the cryptography algorithms is embodied or used in the file, replacing the contents in the file associated with the cryptography algorithm with corresponding contents that are not susceptible to compromise by a quantum computer.