liveBook · Manning

Probabilistic model

As explained in Machines that Think

A mathematical model that represents uncertainty using probability distributions over variables or outcomes.

A probabilistic model represents uncertainty using probability distributions over variables or outcomes. In artificial intelligence, probabilistic models support reasoning when information is incomplete or ambiguous, shifting the emphasis from fixed logic to inference. section 1.7

This approach contrasts with rule-based systems, whose reasoning can be traced through rules written in advance. Some probabilistic models learn parameters from data to capture patterns that designers have not specified explicitly. Their decisions can be harder to interpret because influences are distributed across numerical relationships rather than a single chain of logic. section 1.4.4 section 1.7

That opacity is part of the black box problem: an output may be observable even when the reasons for a particular decision are difficult to explain. This can complicate accountability, safety checks, bias detection, and debugging. Probabilistic models also offer a way to explore the idea that aspects of cognition can emerge from reasoning under uncertainty. section 1.4.4 section 1.6.1

How Machines that Think covers it

How deep it goes
Outlines the shift from rules to inference and discusses opacity, without detailing probability distributions or inference procedures.
What you can do after
Evaluate/understand the trade-off between capturing patterns from data and explaining decisions in complex learned models.
How it connects
Connects probabilistic models with symbolic reasoning, neural networks, inference, and the black box problem.

Also explained in

Compare how 5 books explain probabilistic model →

Related concepts

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