A pushforward measure is the measure obtained by transporting an existing measure along a measurable map. If maps one measurable space into another and is a measure on , its pushforward assigns a measurable set the value . Usually written or , this construction is fundamental in measure theory and probability: the distribution of a random variable is the pushforward of the underlying probability measure.
Definition
Let and be measurable spaces, and let be a measure on . A map is measurable when
The pushforward of under is the measure on defined by
Here
is a preimage, not an inverse function. Consequently, need not be injective, surjective, continuous, or differentiable.
The sigma-algebra is part of the construction. The pushforward is defined on its sets, not automatically on every subset of .
Why this defines a measure
Preimages preserve the empty set and countable unions. They also preserve disjointness: if are pairwise disjoint, then their preimages are pairwise disjoint. Therefore,
Together with , this establishes the measure axioms.
The definition uses preimages because they have these properties regardless of whether is invertible. Simply assigning mass using images of sets would not provide an analogous general construction: images of disjoint sets can overlap, and images of measurable sets need not be measurable.
Basic properties
Preservation of total mass
Since ,
Thus finite measures remain finite, and probability measures remain probability measures.
More generally, the construction is linear for nonnegative combinations of measures:
Composition
If is another measurable map, then
This follows from
The identity map also satisfies
These identities make pushforward compatible with function composition: transporting a measure through two maps in succession gives the same result as transporting it through their composition.
Changes on null sets
If two measurable maps agree outside a measurable set of -measure zero, then
Thus modifying a map almost everywhere does not alter its pushforward.
Integration against a pushforward
The central integration identity is
for every nonnegative measurable function . It also holds for real- or complex-valued integrable functions.
In particular,
so is integrable with respect to exactly when is integrable with respect to .
The identity is first checked for indicator functions:
It then extends to nonnegative simple functions and, by monotone approximation, to nonnegative measurable functions. Integrable functions follow by decomposition.
This is an abstract change-of-variables formula for the Lebesgue integral. It requires no derivatives, Jacobians, or invertibility. Those enter only when expressing the pushforward relative to a particular reference measure.
The identity also shows a useful duality: measures move forward under , while functions are composed backward with .
Probability distributions
Let be a probability space, and let be a random variable or random element. Its probability distribution, also called its law, is
For each measurable ,
If is measurable, composition gives
Thus the distribution of a transformed random variable depends only on its original distribution and the transformation, not on the particular probability space used to realize it.
The integration identity becomes
whenever the expectation is defined.
For real-valued , its cumulative distribution function is
A joint distribution is likewise a pushforward under the map
Its marginal distributions are obtained by pushing that joint measure forward under the coordinate projections.
Examples
Discrete measures
If a finite measure is concentrated on points , with masses , then
Points with the same image contribute to the same target mass.
For example, suppose assigns masses to , respectively. Under , the pushforward assigns mass to and mass to .
A constant map
If and , then
In particular, any probability measure can be pushed forward to a measure concentrated at a single point. Pushforward therefore need not preserve the absence of atoms.
Squaring a uniform random variable
Let have the uniform distribution on , and set . For ,
Consequently, has density
and zero density outside that interval. The singularity at zero is integrable and does not represent an atom.
Densities and Jacobian formulas
A pushforward measure always exists under the measurability assumptions, but it need not have a density. Density formulas require additional hypotheses.
Suppose are open, is a continuously differentiable bijection with continuously differentiable inverse, and
where denotes Lebesgue measure. Then
Equivalently,
Here is the Jacobian matrix, and its determinant describes local volume scaling.
For an invertible affine transformation , this becomes
In one dimension, a non-injective map can often be handled by summing over inverse branches:
For example, this formula holds almost everywhere when the source density is supported on finitely many open intervals on each of which is continuously differentiable with nonzero derivative. Each preimage branch contributes separately.
If a map collapses a positive amount of mass onto a point or a lower-dimensional set, the pushforward can have a singular component instead of an ordinary density.
Limitations and measure-theoretic subtleties
Sigma-finiteness need not be preserved
A measure is sigma-finite if its space is a countable union of sets of finite measure. Even when is sigma-finite, need not be.
For example, let be Lebesgue measure on , and map every real number to a single point. The resulting one-point space has infinite measure, and its only finite-measure measurable set is the empty set. Its measure is therefore not sigma-finite.
Absolute continuity can be lost
If has a density relative to Lebesgue measure, a constant map still sends it to a point mass. Thus the absolute continuity needed for a density is not automatically preserved.
There is, however, a relative preservation property: if , then
Indeed, implies , hence .
Information is generally lost
A non-injective map merges points, so its pushforward generally cannot determine the original measure. For , point masses at and have the same pushforward.
By contrast, if is a bijection and both and are measurable, then
Applications
Measure-preserving dynamics
A measurable transformation preserves a measure precisely when
or equivalently,
for every measurable . This is the invariant-measure condition used in the study of a dynamical system.
Optimal transport
In optimal transport, a map transports a source probability measure to a target probability measure when
The map may be chosen to minimize a cost such as
A deterministic transport map need not exist: a source concentrated at one point cannot be mapped deterministically to a target distributing mass over several points. Transport plans extend the formulation by allowing mass from one source point to be split among multiple destinations.
Weak convergence
Suppose probability measures converge weakly to on a metric space, and is continuous into another metric space. Then
For every bounded continuous function on the target, is bounded and continuous on the source, so
This is the measure-theoretic form of the continuous mapping principle used in probability limit theorems.