SECAI Core · phase 2 of 15
Autoencoder
Small neural networks trained to recreate the records they see during learning
The Explain card
- Plain English
- An autoencoder is a small neural network trained to compress its input and then rebuild it as closely as possible. It learns what "normal" looks like by recreating it.
- Example
- A network monitoring tool trains an autoencoder on normal traffic flows. When a new flow reconstructs badly, the high error flags it as a likely anomaly, such as beaconing to a command server.
- Why it matters
- Autoencoders power anomaly detection, but they only know what they were shown. If attacker traffic was in the training set, it becomes "normal" and slips past.
- Hook
- It learns to copy normal, so anything it cannot copy is suspicious.
Where it sits in the deck
Phase 2: How Models Are Built: Architectures and Learning Mechanics
With the paradigms named, zoom in on the architectural building blocks that turn data into trained artifacts.
