Quasi-Unitary Embedding

• Quasi-unitary embedding is a family of constructions that approximately preserve unitary structure through controlled errors, diagonal forms, or unitisation-compatible mappings.
• The framework uses δ-parameters to compare operators, quadratic forms, and matrices, ensuring that key properties such as Hermitian dual containment are maintained.
• Its versatile applications span spectral convergence, coding theory, and quasi-periodic dynamics, effectively bridging discrete and continuous mathematical models.

“Quasi-unitary embedding” does not denote a single universally fixed notion across the arXiv literature. Instead, it labels a family of constructions in which an embedding preserves unitary structure exactly, approximately, or in a weakened algebraic sense. In one formal usage, it means comparison of operators or quadratic forms on different Hilbert spaces by identification maps that are unitary up to a controlled error parameter $\delta$ (Post et al., 28 Mar 2025). In coding theory, it refers to matrices $B\in \mathcal M(\mathbb F_{q^2},k\times k)$ satisfying $BB^\dagger$ diagonal over $\mathbb F_q^\ast$, which is sufficient to preserve Hermitian dual-containing structure under matrix-product code constructions (Cao, 2020). In dynamical systems, the phrase is used in connection with embedding a quasi-periodic cocycle as the Poincaré map of a nearby analytic quasi-periodic linear flow (You et al., 2012). In quantum-information-inspired geometry, it describes attempts to embed spaces of quregisters into $\mathrm{GL}(2^n)$ or $\mathrm{SU}(2^n)$, a program that succeeds for qubits and fails globally for $n\ge 2$ (Cervantes et al., 2016). This variety suggests that the expression functions less as a single definition than as a recurrent structural motif.

1. Terminological scope and delimitations

Several papers explicitly separate “quasi-unitary embedding” from nearby but distinct notions. In manifold learning, the relevant concept is a quasi-isometric embedding: the projection $p\in\mathcal G(m,k)$ is chosen to minimize the worst secant distortion

$$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$

with $\Sigma$ the set of normalized secants of the sampled point cloud. The paper states that this is “approximate isometry under dimension reduction, not approximate unitarity,” even though both ideas concern controlled deviation from ideal length preservation (Dreisigmeyer, 2017).

A different, unrelated use of “embedding” occurs in geometric group theory. There, a map between finitely generated groups is a quasi-isometric embedding if there exist constants $B\in \mathcal M(\mathbb F_{q^2},k\times k)$0, $B\in \mathcal M(\mathbb F_{q^2},k\times k)$1 such that

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$2

and the theory is applied, for example, to embeddings $B\in \mathcal M(\mathbb F_{q^2},k\times k)$3 between generalized Thompson groups (Sheng, 2018). This is a coarse metric notion rather than a unitary or almost-unitary one.

Other literatures replace the phrase by more precise terms. One paper studies quasi-unitary equivalence of operators and forms (Post et al., 28 Mar 2025); another studies embedding of selfadjoint operator spaces via complete isometry of Werner unitisations and notes that it does not use the exact phrase “quasi-unitary embedding” as its main term (Chatzinikolaou et al., 28 Oct 2025). A plausible implication is that encyclopedia treatment is best organized by mathematical mechanism rather than by a single lexical definition.

2. Quasi-unitary equivalence on varying Hilbert spaces

The most explicit formalization appears in the theory of non-negative self-adjoint operators $B\in \mathcal M(\mathbb F_{q^2},k\times k)$4 on $B\in \mathcal M(\mathbb F_{q^2},k\times k)$5 and $B\in \mathcal M(\mathbb F_{q^2},k\times k)$6 on $B\in \mathcal M(\mathbb F_{q^2},k\times k)$7, compared by identification operators

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$8

The paper says that $B\in \mathcal M(\mathbb F_{q^2},k\times k)$9 is $BB^\dagger$0-quasi-unitary with $BB^\dagger$1-quasi-adjoint $BB^\dagger$2 if

$BB^\dagger$3

and

$BB^\dagger$4

The operators are then $BB^\dagger$5-quasi-unitarily equivalent if, in addition,

$BB^\dagger$6

When $BB^\dagger$7, $BB^\dagger$8 is genuinely unitary, $BB^\dagger$9, and the construction reduces to ordinary unitary equivalence (Post et al., 28 Mar 2025).

The same framework is formulated at the level of closed, non-negative quadratic forms $\mathbb F_q^\ast$0 and $\mathbb F_q^\ast$1. Besides $\mathbb F_q^\ast$2, one introduces

$\mathbb F_q^\ast$3

with $\mathbb F_q^\ast$4 and $\mathbb F_q^\ast$5. The forms are $\mathbb F_q^\ast$6-close if

$\mathbb F_q^\ast$7

This form-level notion is the paper’s fundamental abstract definition, because form domains are often easier to control than operator domains (Post et al., 28 Mar 2025).

Its principal consequence is generalized norm resolvent convergence on varying Hilbert spaces. If $\mathbb F_q^\ast$8 and $\mathbb F_q^\ast$9 are $\mathrm{GL}(2^n)$0-quasi-unitarily equivalent, then

$\mathrm{GL}(2^n)$1

for $\mathrm{GL}(2^n)$2 and $\mathrm{GL}(2^n)$3, with

$\mathrm{GL}(2^n)$4

The theory further yields convergence of spectral projections, heat operators, eigenvalues, and eigenfunctions, and it proves spectral exactness on compact spectral sets. The framework is stable under composition: if $\mathrm{GL}(2^n)$5 and $\mathrm{GL}(2^n)$6 are $\mathrm{GL}(2^n)$7-quasi-unitarily equivalent and $\mathrm{GL}(2^n)$8 and $\mathrm{GL}(2^n)$9 are $\mathrm{SU}(2^n)$0-quasi-unitarily equivalent, then $\mathrm{SU}(2^n)$1 and $\mathrm{SU}(2^n)$2 are $\mathrm{SU}(2^n)$3-quasi-unitarily equivalent at operator level, while the form-level transitivity estimate is $\mathrm{SU}(2^n)$4 (Post et al., 28 Mar 2025).

Two example classes illustrate the abstraction. For graph approximations of the unit interval, the graph energy forms and the continuum energy form are $\mathrm{SU}(2^n)$5-quasi-unitarily equivalent with

$\mathrm{SU}(2^n)$6

For the Sierpiński gasket, the corresponding rate is

$\mathrm{SU}(2^n)$7

The paper also treats manifolds with small obstacles, where the Neumann Laplacian on $\mathrm{SU}(2^n)$8 is compared to the Laplacian on $\mathrm{SU}(2^n)$9 by natural restriction and extension maps (Post et al., 28 Mar 2025).

3. Finite-field quasi-unitary matrices and matrix-product codes

In coding theory, a matrix $n\ge 2$0 is quasi-unitary if

$n\ge 2$1

If $n\ge 2$2, then $n\ge 2$3 is unitary. The essential point is that the Hermitian Gram matrix is diagonal over the base field, not necessarily the identity. The paper emphasizes that this diagonal form is enough to preserve Hermitian dual-containing structure in matrix-product codes (Cao, 2020).

The matrix-product construction is

$n\ge 2$4

where the constituent codes have the same length and $n\ge 2$5. If $n\ge 2$6 has full rank and $n\ge 2$7 are $n\ge 2$8, then

$n\ge 2$9

In the Hermitian setting, the dual satisfies

$p\in\mathcal G(m,k)$0

If $p\in\mathcal G(m,k)$1 is quasi-unitary and $p\in\mathcal G(m,k)$2, then $p\in\mathcal G(m,k)$3 is Hermitian dual-containing, hence yields a quantum code with parameters

$p\in\mathcal G(m,k)$4

This is the Hermitian analogue of the quasi-orthogonal Euclidean construction (Cao, 2020).

The mechanism is algebraically transparent. If

$p\in\mathcal G(m,k)$5

then

$p\in\mathcal G(m,k)$6

Thus the Hermitian dual of the matrix-product code is described by the same defining matrix up to diagonal scaling. Since multiplication by nonzero field elements preserves linear-code containment, dual containment survives the embedding effected by $p\in\mathcal G(m,k)$7 (Cao, 2020).

A stronger condition is NSC (“non-singular by columns”). For square $p\in\mathcal G(m,k)$8 NSC matrices,

$p\in\mathcal G(m,k)$9

so the minimum-distance bound becomes

$$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$0

The paper proves that if all leading principal minors of $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$1 are nonzero, then there exists a lower unitriangular matrix $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$2 such that $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$3 is quasi-unitary; if $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$4 is NSC, then $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$5 is NSC quasi-unitary. It also proves that for every $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$6, there exist $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$7 NSC quasi-unitary matrices over $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$8, and gives an explicit Vandermonde-type example when $$D_\Sigma(p)=\max_{\sigma\in\Sigma}\left|1-\|p^T\sigma\|_2^2\right|,$$9: $\Sigma$0 In this literature, “quasi-unitary embedding” is therefore an embedding operator that enlarges length while preserving Hermitian duality data (Cao, 2020).

4. Embedding quasi-periodic cocycles into analytic flows

A dynamical-systems usage arises from the local embedding theorem for analytic quasi-periodic cocycles. For a frequency vector $\Sigma$1 with $\Sigma$2 rationally independent and a constant matrix $\Sigma$3, the theorem states that if

$\Sigma$4

then the quasi-periodic cocycle

$\Sigma$5

can be analytically embedded into a quasi-periodic linear system. More precisely, there exist $\Sigma$6 and

$\Sigma$7

such that $\Sigma$8 is the Poincaré map of

$\Sigma$9

The main proof is written for $B\in \mathcal M(\mathbb F_{q^2},k\times k)$00, and the theorem is stated more generally for

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$01

(You et al., 2012).

In the cited summary, this embedding phenomenon is described as closely tied to the idea of a quasi-unitary embedding: the cocycle is lifted to a nearby analytic quasi-periodic flow on an extended torus, and the cocycle is recovered as a time-one Poincaré map. The proof uses a linear embedding operator solving a cohomological equation on a resonance-supported Banach subspace, together with an implicit function theorem applied to a functional whose zero set encodes

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$02

A special remark is that if $B\in \mathcal M(\mathbb F_{q^2},k\times k)$03 is diagonalizable in the relevant real normal form, then one can take $B\in \mathcal M(\mathbb F_{q^2},k\times k)$04 (You et al., 2012).

The structural significance is expressed by the equivalence theorem: an analytic quasi-periodic linear system $B\in \mathcal M(\mathbb F_{q^2},k\times k)$05 is almost reducible if and only if its corresponding Poincaré cocycle $B\in \mathcal M(\mathbb F_{q^2},k\times k)$06 is almost reducible; the same equivalence holds for rotations reducibility. The embedding theorem is the bridge that transfers local almost reducibility results from flows to cocycles and global reducibility results from cocycles to flows (You et al., 2012).

The framework also yields spectral consequences. Via Aubry duality and reducibility, the paper proves Anderson localization for the long-range operator

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$07

under the condition

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$08

for irrational $B\in \mathcal M(\mathbb F_{q^2},k\times k)$09 with finite Liouville exponent $B\in \mathcal M(\mathbb F_{q^2},k\times k)$10. For the almost Mathieu operator, the corresponding condition is

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$11

This use of embedding is thus not merely representational; it serves as a transfer principle between discrete and continuous quasi-periodic dynamics (You et al., 2012).

5. Quregisters, special linear groups, and entanglement-sensitive failure of unitarity

A more geometric and quantum-information-oriented formulation studies embeddings of the unit sphere of $B\in \mathcal M(\mathbb F_{q^2},k\times k)$12 into matrix groups. For $B\in \mathcal M(\mathbb F_{q^2},k\times k)$13, the central map is

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$14

Because $B\in \mathcal M(\mathbb F_{q^2},k\times k)$15, the columns are orthonormal and

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$16

so $B\in \mathcal M(\mathbb F_{q^2},k\times k)$17 is a bijection $B\in \mathcal M(\mathbb F_{q^2},k\times k)$18. The group law of $B\in \mathcal M(\mathbb F_{q^2},k\times k)$19 is therefore transported to the qubit sphere by

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$20

(Cervantes et al., 2016).

The paper emphasizes that this picture is special to $B\in \mathcal M(\mathbb F_{q^2},k\times k)$21. For $B\in \mathcal M(\mathbb F_{q^2},k\times k)$22, one has

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$23

and the natural candidate is a tensor-product construction based on $B\in \mathcal M(\mathbb F_{q^2},k\times k)$24. For a $B\in \mathcal M(\mathbb F_{q^2},k\times k)$25-quregister $B\in \mathcal M(\mathbb F_{q^2},k\times k)$26, separability is characterized by

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$27

On the separable locus, the chartwise maps $B\in \mathcal M(\mathbb F_{q^2},k\times k)$28 land in $B\in \mathcal M(\mathbb F_{q^2},k\times k)$29, and the tensor-product construction behaves correctly. Globally, however, the paper states that the embedding obtained by tensoring $B\in \mathcal M(\mathbb F_{q^2},k\times k)$30 fails to determine a bijection between $B\in \mathcal M(\mathbb F_{q^2},k\times k)$31 and $B\in \mathcal M(\mathbb F_{q^2},k\times k)$32 (Cervantes et al., 2016).

This failure is quantified by entanglement. Writing

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$33

the reduced density matrices have eigenvalues

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$34

so the reduced von Neumann entropy is

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$35

Moreover, the paper defines

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$36

and derives

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$37

Thus

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$38

while Bell states are maximally entangled and give $B\in \mathcal M(\mathbb F_{q^2},k\times k)$39. In this setting, quasi-unitary embedding is exact on the separable submanifold and fails on entangled states in a way that becomes an entanglement measure consistent with von Neumann entropy (Cervantes et al., 2016).

6. Operator-algebraic and representation-theoretic analogues

In selfadjoint operator-space theory, the nearest formal analogue is an embedding compatible with Werner’s unitisation. For a completely isometric complete order embedding

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$40

the paper calls $B\in \mathcal M(\mathbb F_{q^2},k\times k)$41 an embedding if the induced map on unitisations

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$42

is completely isometric, equivalently a unital complete order embedding. This condition is characterized by several extension properties: every bounded positive functional on each matrix level extends with the same norm; quasistates extend to quasistates; every completely contractive completely positive map into $B\in \mathcal M(\mathbb F_{q^2},k\times k)$43 or $B\in \mathcal M(\mathbb F_{q^2},k\times k)$44 extends with the same cb-norm; and $B\in \mathcal M(\mathbb F_{q^2},k\times k)$45 is a gauge maximal isometry in Russell’s sense. The key distance formula is

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$46

The paper also notes explicitly that it does not use “quasi-unitary embedding” as its main term, although the unitisation-compatible embedding property is the closest related notion (Chatzinikolaou et al., 28 Oct 2025).

A different operator-algebraic direction is given by unitary correlation sets. With

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$47

the universal C$B\in \mathcal M(\mathbb F_{q^2},k\times k)$48-algebra generated by entries of an abstract unitary matrix, and

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$49

the paper studies tensor-product models $B\in \mathcal M(\mathbb F_{q^2},k\times k)$50, $B\in \mathcal M(\mathbb F_{q^2},k\times k)$51, and compressed sets $B\in \mathcal M(\mathbb F_{q^2},k\times k)$52. Its central equivalence is that Connes’ embedding problem has a positive answer if and only if

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$53

equivalently if and only if

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$54

equivalently if and only if

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$55

isometrically for all $B\in \mathcal M(\mathbb F_{q^2},k\times k)$56. Here the embedding-type question is encoded in finite-dimensional tensor norms rather than in a single explicit map between Hilbert spaces (Harris et al., 2016).

In automorphic representation theory, the analogue is an $B\in \mathcal M(\mathbb F_{q^2},k\times k)$57-group embedding for quasi-split unitary groups. The paper defines

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$58

and

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$59

and constructs canonical $B\in \mathcal M(\mathbb F_{q^2},k\times k)$60-embeddings

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$61

depending on a character $B\in \mathcal M(\mathbb F_{q^2},k\times k)$62. The image is identified with the conjugate self-dual parameters of parity

$B\in \mathcal M(\mathbb F_{q^2},k\times k)$63

Although the paper does not use the phrase “quasi-unitary embedding” explicitly, the entire endoscopic framework is organized around these structural embeddings and the transfer of parameters, packets, and automorphic representations between unitary groups and twisted forms of $B\in \mathcal M(\mathbb F_{q^2},k\times k)$64 (Mok, 2012).

Taken together, these operator-theoretic, coding-theoretic, dynamical, geometric, and representation-theoretic usages show that “quasi-unitary embedding” is best understood as a family of constructions in which exact unitarity is replaced by a controlled defect, a diagonal Gram form, a unitisation-compatible order structure, or a transfer map into a larger unitary framework. The common theme is not literal uniformity of definition, but the preservation of a unitary-type invariant under embedding.