Bayesian networks, d-separation and inference
Research context
postgraduate concept · established · As of 2026-10-10
A Bayesian network uses a directed acyclic graph and local conditional distributions to factor a joint law.
Key knowledge
A Bayesian network uses a directed acyclic graph and local conditional distributions to factor a joint law. d-Separation supplies conditional independences implied by its Markov structure; the converse inference from observed independences needs additional faithfulness assumptions. Variable elimination sums latent variables exactly, but cost depends exponentially on induced width in general. The DAG alone does not encode a causal intervention without additional causal assumptions.
Values and equations
p(x_1,...,x_m)=product_i p(x_i|x_pa(i))
- pa(i)
- Parents of node i in a directed acyclic graph
- p
- Joint probability mass/density compatible with the graph
Conditions and scope
- The graph is acyclic and distributions/parameters define a normalized joint law.
- Conditional independence is not marginal independence; conditioning on a collider can induce association.
Sources · 2
References
Sources retain their recorded scope. A source link alone does not establish verification of every statement.
Evidence and review
Author checked this English statement, displayed equation where present and stated hypotheses against these source locations: DAG factorization, d-separation, interventions, back-door adjustment and identifiability; §§2–4: KL/ELBO, mean field, coordinate ascent and MCMC comparisons. The check is bounded to the recorded claims; no complete proof reconstruction, software proof replay or universal literature completeness is claimed.
2026-10-10 · Author source-body/statement/equation/domain check, with graph associations scoped individually. Separate independent content/source/inventory review is required before branch closure.
Entry details
Permanent link to this entry ↗- Record identity
- research-info-bayesian-network-factorization
- Revision
- research.1
- Structured reference record
- Postgraduate & research reference