Brenier map

E1017920

The Brenier map is the unique gradient of a convex function that provides the optimal transport between probability measures under a quadratic cost, playing a central role in modern optimal transport theory.

Try in SPARQL Jump to: Statements Referenced by

Statements (47)

Predicate Object
instanceOf mathematical concept
optimal transport map
appearsIn polar factorization of vector-valued maps
hasCodomain Euclidean space NERFINISHED
hasDomain Euclidean space NERFINISHED
hasRegularityDependingOn regularity of source and target densities GENERATED
is almost everywhere uniquely defined
gradient of a convex function
optimal transport map for quadratic cost
isAssociatedWith L2 optimal transport
Wasserstein-2 distance
isCharacterizedAs Monge solution of the quadratic optimal transport problem
gradient of a convex potential
isDefinedBetween absolutely continuous probability measures
probability measures
isDefinedUnder quadratic cost
isGeneralizedBy optimal maps for other cost functions
isGradientOf convex potential function
isGuaranteedBy Brenier’s polar factorization theorem NERFINISHED
isPushforwardMapOf convex potential gradient
isRelatedTo Kantorovich formulation of optimal transport NERFINISHED
Monge formulation of optimal transport NERFINISHED
cyclical monotonicity
isSolutionOf Monge problem with quadratic cost
isSpecialCaseOf c-convex potential gradients for c(x,y)=|x−y|^2/2
isStableUnder weak convergence of measures (under suitable conditions)
isSubjectOf regularity theory in optimal transport
isToolIn fluid mechanics
geometric analysis
machine learning
modern optimal transport theory
partial differential equations
probability theory
isUnique up to source-measure null sets
isUsedFor density equalization mappings
domain morphing and mesh transport
image registration in imaging sciences
measure-preserving rearrangements
isUsedToDefine displacement interpolation
geodesics in Wasserstein-2 space
minimizes quadratic transport cost
namedAfter Yann Brenier NERFINISHED
pushesForward source measure to target measure
requiresAssumption source measure absolutely continuous with respect to Lebesgue measure
satisfies Monge–Ampère equation in suitable settings
cyclical monotonicity of its graph
underlies Riemannian structure of Wasserstein space

Referenced by (2)

Full triples — surface form annotated when it differs from this entity's canonical label.