4.13 Bayesian networks: representation and inference
You can read and build a Bayesian network, answer queries by exact inference, and say when a probabilistic representation is the right one.
Before:01. Mathematics for Machine Learning02. Python — Basics to Advanced
A Bayesian network is a picture of conditional independence — the graph is the assertion, the tables are the detail — and its value is compactness: a joint distribution that would need thousands of numbers fits in a handful. This topic covers construction from a domain description and query answering by exact inference, including where exactness stops being feasible. It sits here as the probabilistic member of the representation family, and the closing judgement — when probabilistic beats symbolic — is the module's question in miniature.
Work through these
Represent conditional independence as a graph, and state what a network assumes
A graph where an arrow means direct influence, encoding which variables become independent once others are known. The graph is a compact statement of assumptions rather than a picture of data.
Construct a network from a domain description, conditional probability tables included
Building a network from a described situation, including the tables that say how likely each variable is given its parents. Getting the arrow directions right is a modelling decision with real consequences.
Introduction to Probability & Statistics (18.05) · CourseRead a conditional probability table correctly
Reading a conditional probability table correctly means being clear about what is being conditioned on. Misreading the direction is the most common error in this topic.
Introduction to Probability & Statistics (18.05) · CourseAnswer a query by exact inference, and say when that stops being feasible
Computing the probability of one variable given evidence about others, exactly, and knowing that exact computation becomes infeasible as networks grow. That limit is why approximate methods exist.
Judge when a probabilistic representation beats a symbolic one
A probabilistic representation handles uncertainty and conflicting evidence that a purely symbolic one cannot. Saying when that advantage is worth its cost is the judgement being taught.
Sign in to keep your progress.
Free resources
We haven't checked most of these for screen reader use yet.
Links last checked 29 Aug 2026.
Stuck here?
Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.
Checking your session…
Topics shown in module order.