Wireless spectrum anomaly detector
Non-documented SAFE base code used for the following papers
- SAIFE: Unsupervised wireless spectrum anomaly detection with interpretable features
- Unsupervised Wireless Spectrum Anomaly Detection With Interpretable Features
- Crowdsourced wireless spectrum anomaly detection
Adapted from: https://github.com/hwalsuklee/tensorflow-mnist-AAE . Thanks to musyoku's pytorch implementaion for all the unsupervised setup and quick answers (https://github.com/musyoku/adversarial-autoencoder).
Requirements: Tested on
Python 3.6.9, Tensorflow 1.12.0, Tflearn
Sample run:
python spectrum_semsup_cat.py --prior_type normal --num_epochs=510 --PMLR_n_samples=500 --learn_rate=0.05e-3 --dimz=50
More documentation and examples will be updated soon.