Interval#
- class hybrid_aif.Interval[source]#
Bases:
ModuleA closed interval [lower, upper], broadcastable like its member arrays.
The two flags are static fields, so they cannot be traced and must agree across a
lax.scancarry; each combinator below propagates them.- derived: bool = True#
Whether the enclosure is guaranteed, given a correct implementation, on the model the planner was handed;
Falsemarks a number a net produced, which nothing bounds. A result is derived iff every input is (an AND through every combinator). Estimators declare it once at registration (hybrid_aif.objectives.registry).
- approximate: bool = False#
Whether that model is a linearised or quadrature reading of the declared one (a nonlinear transition or a neural likelihood), so the enclosure holds for the reading, not for the declaration. ORed through every combinator: a result is approximate iff any input is.
- classmethod exact(value, derived=True, approximate=False)[source]#
A zero-width interval: the estimate is a point.
derived=Falseis the amortized leaf, a point estimate with no warrant (a critic’s value, an ensemble’s Monte-Carlo info gain).approximate=Truemarks a point read off a linearised or quadrature model.
- __init__(lower, upper, derived=True, approximate=False)#