Machine Learning Model Score Obfuscation Using Vector Modification Techniques

Patent metadata

Field	Value
Publication number	US-2020349401-A1
Application number	US-201916399701-A
Country	US
Kind code	A1
Filing date	Apr 30, 2019
Priority date	Apr 30, 2019
Publication date	Nov 5, 2020
Grant date	—

Abstract

An artefact is received. Features from such artefact are extracted and then populated in a vector. Subsequently, one of a plurality of available dimension reduction techniques are selected. Using the selected dimension reduction technique, the features in the vector are reduced. The vector is then input into a classification model and the score can be provided to a consuming application or process. Related apparatus, systems, techniques and articles are also described.

First claim

What is claimed is:

1 . A computer-implemented method comprising:
receiving an artefact; extracting features from the artefact and populating a vector; selecting one of a plurality of available dimension reduction techniques; reducing the features in the vector using the selected dimension reduction technique; inputting the vector into a classification model to generate a score; and providing the score to a consuming application or process.

2 . The method of claim 1 , wherein the classification model characterizes the artefact as being malicious or benign to access, execute, or continue to execute.

3 . The method of claim 2 further comprising: preventing access or execution of the artefact when the classification model characterizes the artefact as being malicious.

4 . The method of claim 1 , wherein the available dimension reduction technique is selected using a random selection algorithm.

5 . The method claim 1 , wherein the available dimension reduction technique is selected using a load balancing algorithm.

6 . The method of claim 1 , wherein the available dimension reduction technique is selected using a round robin selection algorithm.

7 . The method of claim 1 further comprising training the classification model to take into account the use of multiple feature reduction techniques.

8 . The method of claim 1 , wherein at least one of the dimension reduction techniques utilizes principal component analysis.

9 . The method of claim 1 , wherein at least one of the dimension reduction techniques utilizes random projection matrices.

10 . The method of claim 1 , wherein the classification model is a machine learning model trained using a training data set and providing a continuous scale output.

11 . The method of claim 10 , wherein the machine learning model comprises one or more of: a logistic regression model, a neural network, a concurrent neural network, a recurrent neural network, a generative adversarial network, a support vector machine, a random forest, or a Bayesian model.

11 . A computer-implemented method comprising:
receiving an artefact; extracting features from the artefact and populating a vector; fuzzing the vector using an obfuscation algorithm; inputting the vector into a classification model to generate a score; and providing the score to a consuming application or process.

12 . The method of claim 11 , wherein the classification model characterizes the artefact as being malicious or benign to access, execute, or continue to execute.

13 . The method of claim 12 further comprising: preventing access or execution of the artefact when the classification model characterizes the artefact as being malicious.

14 . The method of claim 1 , further comprising:
reducing the features in the vector using a dimension reduction technique

15 . The method of claim 14 , wherein the dimension reduction technique utilizes principal component analysis.

16 . The method of claim 14 , wherein the dimension reduction technique utilizes random projection matrices.

17 . The method of claim 11 , wherein the classification model is a machine learning model trained using a training data set and providing a continuous scale output.

18 . The method of claim 17 , wherein the machine learning model comprises one or more of: a logistic regression model, a neural network, a concurrent neural network, a recurrent neural network, a generative adversarial network, a support vector machine, a random forest, or a Bayesian model.

19 . The method of claim 11 further comprising training the classification model to take into account the fuzzing of the vector using the obfuscation algorithm.

20 . A system comprising:
at least one data processor; and memory storing instructions which, when executed by the at least one data processor, result in operations comprising:
receiving an artefact;
extracting features from the artefact and populating a vector;
selecting one of a plurality of available dimension reduction techniques;
reducing the features in the vector using the selected dimension reduction technique;
inputting the vector into a classification model to generate a score; and
providing the score to a consuming application or process.

Assignees

• Cylance Inc

Inventors

• Beveridge David N
• Buckingham Hailey

Classifications

• G06F21/14 Primary

  against software analysis or reverse engineering, e.g. by obfuscation · CPC title

• G06F18/2411

  based on the proximity to a decision surface, e.g. support vector machines · CPC title

• G06F18/2115

  by evaluating different subsets according to an optimisation criterion, e.g. class separability, forward selection or backward elimination · CPC title

• G06F18/24133

  Distances to prototypes · CPC title

• G06F18/2193

  based on specific statistical tests · CPC title