pretrained_models_prerelease

The current repository contains all the pretrained models for each scenario and each Deep Learning (DL) and Machine Learning (ML) algorithms.

Scenarios

The current repository contains all the pretrained models for each scenario and each Deep Learning (DL) and Machine Learning (ML) algorithms.

• type classification
• Family classification
• Novelty classification
• Virtualization identification
• Packer identification
• Obfuscation classification
• Executable classification

Accumulators

• Contained accumulated data (ACC) needed for the NICV computation. There is two files per executable:
  • the sum,
  • the sum of the square.

Machine Learning Algorithms

• Naïve Bayesian (NB):
  • 7 models
• Support Vector Machines (SVM):
  • 7 models
• Linear Discriminant Analysis (LDA):
  • 9 models

Deep Learning Algorithms

• Convolutional Neural Network (CNN):
  • 7 models (compressed with 7z)
  • need to be uncompressed before usable with other moduls, run_decompression.sh decompresses files using 7z
• Multilayer Perceptron (MLP):
  • 7 models (compressed with 7z)
  • need to be uncompressed before usable with other moduls, run_decompression.sh decompresses files using 7z