liveBook · Manning

Inference engine

As explained in AI Governance

An inference engine is software that loads and serves trained machine-learning models to generate predictions or outputs.

Inference engines are components in GenAI systems, alongside plugins, orchestration frameworks, and vector databases. A vulnerable or malicious engine can become a trusted entry point to data or networks; compromising surrounding software can undermine a system without attacking its model. section 4.3.3

How AI Governance covers it

How deep it goes
Names inference engines as supply-chain components but gives no engine-specific implementation details or controls.
What you can do after
Understand how a compromised inference engine could provide trusted access to data or networks without attacking the model.
How it connects
Links inference engines to the GenAI supply chain, plugins, orchestration frameworks, and vector databases.

Also explained in

Compare how 5 books explain inference engine →

Related concepts

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