https://github.com/sri-csl/trinity-packet

Code for model safeguards for packet-level DNN classifiers
https://github.com/sri-csl/trinity-packet

Last synced: about 2 months ago
JSON representation

Code for model safeguards for packet-level DNN classifiers

• Host: GitHub
• URL: https://github.com/sri-csl/trinity-packet
• Owner: SRI-CSL
• License: mit
• Created: 2023-12-01T20:36:14.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2023-12-01T20:39:36.000Z (about 1 year ago)
• Last Synced: 2023-12-01T21:33:59.252Z (about 1 year ago)
• Language: Python
• Size: 0 Bytes
• Stars: 0
• Watchers: 18
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
## Context PCAP

cnn.py: Discriminative CNN model for payload classification.

dataset.py: 

    ContextPCAPDataset: class for PCAP dataset (CIC, UNSW, etc.). Inherits from data_structures/dataset.py:NetflowDataset.

    ContextPCAPTorchDataset: class for PyTorch DataLoader. 

experiment.py: 

    ContextPCAPExperiment: class for experiment (Baseline, OOD, etc.). Inherits from data_structures/dataset.py:NetflowExperiment.

fnn.py: Descriminative FNN model for payload classification.

preproces.py: Functions to transform PCAP + flow files into usable datasets. 

transformer.py: Descriminative Transformer model for payload classification.

## Data Structures

dataset.py:

    NetflowDataset: parent class for othery types of datasets (ContextPCAP, SequencePCAP, etc.).

experiment.py:

    NetflowExperiment: parent class that contains functions for training, inference, OOD experiments.

network_model.py:

    NetworkModel: wrapper class for training/inference/feature extraction.

util.py: Miscellaneous parsing functions.

## Detectors

iaf.py:

    IAFDataset: A normalizing flows implementation that allows for model distributions.