Apparatus and methods to classify malware with explainability with artificial intelligence models
Apparatus, systems, and methods to classify malware with explainability are disclosed. An example apparatus includes at least one memory; instructions in the apparatus; and processor circuitry. The example processor circuitry is to execute the instructions to: generate feature vectors from a first input; train a neural network model using a first portion of the feature vectors; add one or more fully connected layers to the trained neural network model to form a hybrid model; validate the hybrid model using a second portion of the feature vectors; and deploy the validated hybrid model as a malware classifier, the malware classifier to provide a malware classification with explainability in response to a second input.
1. An apparatus to enable a malware classification with explainability, the apparatus comprising:
memory circuitry;
instructions; and
processor circuitry to execute the instructions to:
generate feature vectors from a first input;
train a neural network model using a first portion of the feature vectors;
add one or more fully connected layers to the trained neural network model to form a hybrid model;
validate the hybrid model using a second portion of the feature vectors; and
deploy the validated hybrid model as a malware classifier, the malware classifier to provide the malware classification with explainability in response to a second input.
2. The apparatus of claim 1 , wherein the processor circuitry is to remove the one or more fully connected layers to expose new outputs.
3. The apparatus of claim 1 , wherein the one or more fully connected layers form a multi-layer perceptron network.
4. The apparatus of claim 1 , wherein the neural network model includes a convolutional neural network model with max pooling.
5. The apparatus of claim 1 , wherein the feature vectors include at least one of malware identification rules or threat techniques.
6. The apparatus of claim 1 , wherein the feature vectors include static features extracted from a portable executable.
7. The apparatus of claim 1 , wherein the processor circuitry is to update the hybrid model to deploy an updated malware classifier based on at least one of feedback and new input.
8. The apparatus of claim 1 , wherein the malware classifier is to provide explainability by indicating a portion of the second input resulting in the malware classification.
9. The apparatus of claim 8 , wherein the second input includes a string feature of at least one of an executable or software code.
10. A non-transitory computer readable storage medium comprising instructions which, when executed, cause at least one processor to at least:
generate feature vectors from a first input;
train a neural network model using a first portion of the feature vectors;
add one or more fully connected layers to the trained neural network model to form a hybrid model;
validate the hybrid model using a second portion of the feature vectors; and
deploy the validated hybrid model as a malware classifier, the malware classifier to provide a malware classification with explainability in response to a second input.
11. The non-transitory computer readable storage medium of claim 10 , wherein the instructions, when executed, cause the at least one processor to remove the one or more fully connected layers after validating the hybrid model.
12. The non-transitory computer readable storage medium of claim 10 , wherein the first input includes a portable executable and wherein the instructions, when executed, cause the at least one processor to extract static features from the portable executable.
13. The non-transitory computer readable storage medium of claim 10 , wherein the instructions, when executed, cause the at least one processor to update the hybrid model to deploy an updated malware classifier based on at least one of feedback and new input.
14. The non-transitory computer readable storage medium of claim 10 , wherein the instructions, when executed, cause the at least one processor to provide explainability by indicating a portion of the second input resulting in the malware classification.
15. A method to enable malware classification with explainability, the method comprising:
generating, by executing an instruction with processor circuitry, feature vectors from a first input;
training, by executing an instruction with the processor circuitry, a neural network model using a first portion of the feature vectors;
adding, by executing an instruction with the processor circuitry, one or more fully connected layers to the trained neural network model to form a hybrid model;
validating, by executing an instruction with the processor circuitry, the hybrid model using a second portion of the feature vectors; and
deploying, by executing an instruction with the processor circuitry, the validated hybrid model as a malware classifier, the malware classifier to provide a malware classification with explainability in response to a second input.
16. The method of claim 15 , further including removing the one or more fully connected layers.
17. The method of claim 15 , wherein the first input includes a portable executable and further including extracting static features from the portable executable.
18. The method of claim 15 , further including updating the hybrid model to deploy an updated malware classifier based on at least one of feedback and new input.
19. The method of claim 15 , wherein explainability is provided by indicating a portion of the second input resulting in the malware classification.
20. An apparatus comprising:
means for generating feature vectors from a first input;
means for training a neural network model using a first portion of the feature vectors, the means for training to add one or more fully connected layers to the trained neural network model to form a hybrid model and validate the hybrid model using a second portion of the feature vectors; and
means for deploying the validated hybrid model as a malware classifier, the malware classifier to provide a malware classification with explainability in response to a second input.