Where does blanket collide?

In a Bayesian or probabilistic graphical model, a node's Markov blanket is the minimal set of other nodes that renders it conditionally independent of the rest of the network: its parents, its children, and its children's other parents. Given the values of its blanket, no other variable adds information about the node. Pearl's 1988 Probabilistic reasoning in intelligent systems introduced and named the construct within the theory of belief networks. The referent is a conditional-independence boundary in a graph that localizes inference and learning, and it has no connection to a rubber press cylinder or a toddler's blanket.[1]

Unit

The parents, children, and co-parents that make a node conditionally independent

Recovery means

Not applicable

Measured by

Identified from the graph structure as the set beyond which a node is conditionally independent of all others

In developmental psychology, a security blanket is a everyday instance of what Winnicott called the transitional object: a soft possession such as a blanket or soft toy that a young child invests with intense attachment and uses to self-soothe, bridging the move from merger with the caregiver toward independent coping with the world. Winnicott's Playing and reality set out the transitional object and transitional phenomena as a normal developmental stage, not a sign of disturbance. The referent is a self-soothing attachment object and the psychological function it serves, unrelated to a press blanket or a network's shielding variables.[2]

Unit

A child's attachment to and self-soothing use of a comfort object

Recovery means

Not applicable

Measured by

Assessed clinically and observationally by the presence, strength, and regulatory use of attachment to a transitional object