liveBook · Manning

VAEGMM

As explained in Outlier Detection in Python

VAEGMM is a model that combines a variational autoencoder with a Gaussian mixture model for representation learning and outlier detection.

VAEGMM combines a variational autoencoder with a Gaussian mixture model for representation learning and outlier detection. The encoder creates a low-dimensional embedding, while the mixture model evaluates that embedding for unusual observations. section 16.4

The detection pipeline has two linked stages:

  • Encode inputs into a low-dimensional representation. section 16.4
  • Evaluate the learned embedding with a Gaussian mixture model. section 16.4
  • Train the representation model and mixture model together through a shared objective rather than fitting them independently. section 16.4

Use VAEGMM through alibi-detect, which supports deep-learning detectors for tabular, image, and time-series data. Installation requires alibi-detect plus TensorFlow or PyTorch; the API offers flexibility but requires more coding than PyOD. section 16.4

pip install alibi-detect

Install alibi-detect before configuring a VAEGMM-based detector. section 16.4

How Outlier Detection in Python covers it

How deep it goes
Explains the joint encoder–mixture pipeline but stops short of providing VAEGMM-specific implementation code.
Examples
Uses CIFAR10 reconstruction and random-noise perturbation to illustrate related autoencoder outlier detection.
What it assumes
Assumes TensorFlow or PyTorch availability and familiarity with Gaussian mixture modeling.
What you can do after
Apply it via alibi-detect after installing a supported deep-learning framework.
How it connects
Connects variational autoencoders, Gaussian mixture models, representation learning, and outlier detection.
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.