liveBook · Manning

Adversarially learned anomaly detection

As explained in Outlier Detection in Python

Adversarially learned anomaly detection is a GAN-based outlier-detection method using a bidirectional GAN architecture and reconstruction error scoring.

Adversarially learned anomaly detection (ALAD) detects outliers with a GAN-based bidirectional GAN (BiGAN) architecture and reconstruction-error scoring. section 16.2.3

During prediction, an input is encoded, passed through the generator and discriminator, and evaluated through reconstruction error and discriminator probabilities. BiGAN training aligns generated data with real data and latent codes with encoded inputs.

PyOD usage

Create PyOD's ALAD detector, fit it to X, read det.decision_scores_, and optionally inspect learning curves with det.plot_learning_curves().

How Outlier Detection in Python covers it

How deep it goes
Explains BiGAN training, prediction flow, reconstruction error, and discriminator probabilities without deriving the optimization procedure.
Examples
Uses a PyOD ALAD fitting example with decision scores and learning-curve inspection.
What you can do after
Apply it via PyOD's ALAD detector to fit data and inspect outlier decision scores.
How it connects
Connects GANs, BiGANs, latent space, reconstruction error, discriminators, and AnoGAN.
Cover of Outlier Detection in Python
liveBookmanning.com
Read Outlier Detection in Python on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Outlier Detection in Python. The section links point to the parts of the book it draws on.