Hadamard product (matrices)¶

The entrywise product of two matrices of identical shape, multiplying corresponding entries without summing across indices.

Version: v2 · 2026-09-19 · History
Domain-specific #: 4803
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomain: Linear Algebra → Mathematics

Core Idea¶

For equally sized matrices A and B, their Hadamard product A∘B has entry (A∘B)_ij=A_ij B_ij. Index alignment pairs coordinates independently, so scalar multiplication laws lift componentwise and the operation is commutative over commutative scalars. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

The load-bearing residual is not the broad topic of linear algebra. It is It is not ordinary matrix multiplication, tensor product, or composition; unequal shapes require a separately declared broadcasting convention..

How would you explain it like I'm…

Spot-by-Spot Multiplying

Take two egg cartons the same size, each with a number in every cup. Make a new carton by multiplying the number in each cup by the number in the very same cup of the other carton. That spot-by-spot multiplying is the Hadamard product.

Matching-Box Multiplication

A matrix is a grid of numbers. The Hadamard product takes two grids of exactly the same shape and multiplies them spot by spot: top-left times top-left, the next spot times the next spot, and so on. Each answer uses only the two numbers in that one matching spot. Because of this, the order doesn't matter, just like 3×4 equals 4×3. It is different from the usual 'matrix multiplication' that mixes rows and columns together.

Entrywise Matrix Product

For two matrices A and B of the same shape, the Hadamard product A∘B is the matrix whose entry in row i, column j is A_ij times B_ij. Each output entry depends only on the matching pair of input entries, so ordinary rules of number multiplication carry over position by position; for example, it is commutative when the numbers commute. This is not standard matrix multiplication, which mixes rows with columns and needs the inner dimensions to match instead. It is also not a tensor product or function composition. If the shapes differ, it isn't defined unless you separately state a broadcasting rule, as some programming libraries do.

For matrices A and B of identical shape m×n, the Hadamard (entrywise, or Schur) product is defined by (A∘B)_ij = A_ij B_ij. Because indices are aligned and each coordinate is handled independently, scalar algebraic laws (associativity, distributivity over addition, commutativity when the scalars commute) lift componentwise. The identity element is the all-ones matrix of the same shape, not the identity matrix. It must be distinguished from the standard matrix product, which contracts rows against columns, from the Kronecker/tensor product, which enlarges dimensions, and from composition of linear maps. Unequal shapes are not covered by the definition; numerical software may broadcast, but that is an additional convention that must be declared. The test for whether something is a Hadamard product is simply: same shape, and each output entry uses only its matching input pair.

Scope of Application¶

Hadamard product (matrices) belongs to linear algebra and is useful where the analyst can specify two matrices over a compatible scalar algebra, equal dimensions, matched row and column indices, entrywise multiplication, and algebraic properties, then evaluate the factors have the same shape and each output entry uses only the corresponding pair of inputs. The scope is broad within that domain but bounded by the need for the factors have the same shape and each output entry uses only the corresponding pair of inputs. The entry records a descriptive analytical identity; practical use requires the governing domain's evidence, standards, and safety obligations.

Clarity¶

The abstraction clarifies a crowded vocabulary by making the factors have the same shape and each output entry uses only the corresponding pair of inputs the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test. A bare label is insufficient because the name Hadamard product (matrices) can be used for a formal identity, an implementation, or a neighboring result unless carrier and convention are stated.

Manages Complexity¶

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Hadamard product (matrices). Hadamard product (matrices) compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning¶

Identify the carrier. State what the elements, states, objects, or observations are: two matrices over a compatible scalar algebra, equal dimensions, matched row and column indices, entrywise multiplication, and algebraic properties. Reject examples whose alleged carrier belongs to a different problem.
Lock the constitutive rule. Express the factors have the same shape and each output entry uses only the corresponding pair of inputs independently of one notation or implementation.

Knowledge Transfer¶

Knowledge transfers strongly among subfields of linear algebra because they reuse two matrices over a compatible scalar algebra, equal dimensions, matched row and column indices, entrywise multiplication, and algebraic properties, Index alignment pairs coordinates independently, so scalar multiplication laws lift componentwise and the operation is commutative over commutative scalars., and type the carrier, state every parameter and convention in the definition, test that the factors have the same shape and each output entry uses only the corresponding pair of inputs, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions¶

Current abstraction Hadamard product (matrices) Domain-specific

Parents (1) — outgoing typed relations

Hadamard product (matrices) is a kind of Composition Prime

What makes it its own entry: It is not ordinary matrix multiplication, tensor product, or composition; unequal shapes require a separately declared broadcasting convention.

Hierarchy path (1) — routes to 1 parentless root

Hadamard product (matrices) → is a kind of Composition → presupposes Gestalt Principles → is a decomposition of Holism

Neighborhood in Abstraction Space¶

Hadamard product (matrices) sits in a crowded region of the domain-specific corpus (7^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Structured Matrices & Linear Maps (43 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25