OctoInformBeta

OpenAI preprint, September 2026 · 7. Ordinary damping and parameter boundaries · Section 7

Corollary 7.1: Kraus formulas and direct specialization: The Kraus matrices and information formulas specialize at zero stationary excitation

Ordinary damping from the general channel formulas.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

For a generalized amplitude-damping channel with damping strength γ∈[0,1]\gamma\in[0,1] and stationary excited-state population ν∈[0,1]\nu\in[0,1], ordinary amplitude damping is the boundary ν=0\nu=0.

The computational basis of nn qubits consists of bit strings i∈Fin⁡(n)→Fin⁡(2)i\in\operatorname{Fin}(n)\to\operatorname{Fin}(2), where Fin⁡(m)={0,…,m−1}\operatorname{Fin}(m)=\{0,\ldots,m-1\}.✦ An nn-qubit operator is a complex matrix indexed by these bit strings.✦ A density operator ρ\rho is a positive semidefinite operator with Tr⁡ρ=1\operatorname{Tr}\rho=1.✦

Write Kr(γ,ν)K_r(\gamma,\nu), with r∈{0,1,2,3}r\in\{0,1,2,3\}, for the single-qubit Kraus operators. For a label r∈Fin⁡(n)→Fin⁡(4)r\in\operatorname{Fin}(n)\to\operatorname{Fin}(4) and computational basis bit strings ii and jj, the tensor product Kraus operator Kr(n)K_r^{(n)} has the matrix entries in (1): Kr(n)(i,j)=∏k=0n−1Kr(k)(γ,ν)i(k),j(k).K_r^{(n)}(i,j)=\prod_{k=0}^{n-1}K_{r(k)}(\gamma,\nu)_{i(k),j(k)}.(1)✦ For an nn-qubit operator PP, the memoryless channel 𝒩γ,ν⊗n\mathcal N_{\gamma,\nu}^{\otimes n} acts as in (2); it is the tensor power of one fixed single-qubit channel. 𝒩γ,ν⊗n(P)=∑r∈Fin⁡(n)→Fin⁡(4)Kr(n)P(Kr(n))*.\mathcal N_{\gamma,\nu}^{\otimes n}(P) =\sum_{r\in\operatorname{Fin}(n)\to\operatorname{Fin}(4)} K_r^{(n)}P\bigl(K_r^{(n)}\bigr)^*.(2)✦ In (2), ** denotes the adjoint.

Entropy and Classical Information

For a finite complex square matrix PP, the entropy S(P)S(P) is the real part of the trace of the spectral function x↦−xln⁡xx\mapsto-x\ln x applied to PP, with 0ln⁡0=00\ln0=0.✦ For density operators it is the von Neumann entropy in (3), which also specifies the binary entropy hh, both in nats. S(ρ)=−Tr⁡(ρln⁡ρ),h(q)=−qln⁡q−(1−q)ln⁡(1−q),q∈[0,1].S(\rho)=-\operatorname{Tr}(\rho\ln\rho), \qquad h(q)=-q\ln q-(1-q)\ln(1-q),\quad q\in[0,1].(3) The qubit entropy gg as a function of determinant uu is given by (4): g(u)=h(1+1−4u2).g(u)=h\!\left(\frac{1+\sqrt{1-4u}}{2}\right).(4)✦ For input excitation population p∈[0,1]p\in[0,1], the phase-pair output determinant vv and Holevo objective ff are the functions in (5): v(γ,ν,p)=γν(1−ν)+γ(1−γ)(p−ν)2,f(γ,ν,p)=h((1−γ)p+γν)−g(v(γ,ν,p)).\begin{aligned} v(\gamma,\nu,p) &=\gamma\nu(1-\nu)+\gamma(1-\gamma)(p-\nu)^2,\\ f(\gamma,\nu,p) &=h((1-\gamma)p+\gamma\nu)-g(v(\gamma,\nu,p)). \end{aligned}(5) An optimal excitation population pp is a point of [0,1][0,1] satisfying (6): f(γ,ν,q)≤f(γ,ν,p)for every q∈[0,1].f(\gamma,\nu,q)\le f(\gamma,\nu,p) \qquad\text{for every }q\in[0,1].(6)✦

A finite quantum ensemble EE on nn qubits consists of density operators ρa\rho_a and nonnegative weights waw_a summing to one, indexed by a∈{0,…,m−1}a\in\{0,\ldots,m-1\}.✦ The ensemble Holevo information χ(E;𝒩)\chi(E;\mathcal N) of a finite quantum ensemble EE through 𝒩=𝒩γ,ν⊗n\mathcal N=\mathcal N_{\gamma,\nu}^{\otimes n} is the quantity in (7), measured in bits: χ(E;𝒩)=1ln⁡2[S(∑awa𝒩(ρa))−∑awaS(𝒩(ρa))].\chi(E;\mathcal N)=\frac{1}{\ln2} \left[S\!\left(\sum_a w_a\mathcal N(\rho_a)\right) -\sum_a w_a S(\mathcal N(\rho_a))\right].(7)✦ The nn-use Holevo information χ(𝒩γ,ν⊗n)\chi(\mathcal N_{\gamma,\nu}^{\otimes n}) of the channel 𝒩γ,ν⊗n\mathcal N_{\gamma,\nu}^{\otimes n} is the supremum of ensemble Holevo information over finite quantum ensembles on nn qubits, as in (8): χ(𝒩γ,ν⊗n)=supEχ(E;𝒩γ,ν⊗n).\chi(\mathcal N_{\gamma,\nu}^{\otimes n}) =\sup_E\chi(E;\mathcal N_{\gamma,\nu}^{\otimes n}).(8)✦

A classical communication code 𝒞\mathcal C for nn channel uses consists of M≥1M\ge1 encoding density operators ρa\rho_a and a collective decoding positive operator-valued measure {Da}a=0M−1\{D_a\}_{a=0}^{M-1}, with Da≥0D_a\ge0 and ∑aDa=I\sum_aD_a=I.✦ The average decoding error e(𝒞)e(\mathcal C) of a classical communication code 𝒞\mathcal C for equally likely messages is the quantity in (9): e(𝒞)=1−1M∑a=0M−1Re⁡Tr⁡(Da𝒩γ,ν⊗n(ρa)).e(\mathcal C)=1-\frac1M\sum_{a=0}^{M-1}\operatorname{Re}\operatorname{Tr} \!\left(D_a\mathcal N_{\gamma,\nu}^{\otimes n}(\rho_a)\right).(9)✦ A rate RR in bits per channel use is achievable if there is a sequence of codes 𝒞n\mathcal C_n with e(𝒞n)→0e(\mathcal C_n)\to0 such that, for every ε>0\varepsilon>0, their message counts MnM_n satisfy ln⁡Mn/(nln⁡2)≥R−ε\ln M_n/(n\ln2)\ge R-\varepsilon for all sufficiently large nn.✦ The unassisted classical capacity C(𝒩γ,ν)C(\mathcal N_{\gamma,\nu}) of the channel 𝒩γ,ν\mathcal N_{\gamma,\nu} is the supremum of its achievable rates, as in (10): C(𝒩γ,ν)=sup⁡{R∈ℝ:R is achievable}.C(\mathcal N_{\gamma,\nu})=\sup\{R\in\mathbb R:R\text{ is achievable}\}.(10)✦

The Claim

Theorem 1. For real γ\gamma and ν\nu, the single-qubit Kraus operators satisfy (11), (12), (13) and (14): K0(γ,ν)=(1−ν001−ν1−γ).K_0(\gamma,\nu)= \begin{pmatrix} \sqrt{1-\nu}&0\\ 0&\sqrt{1-\nu}\sqrt{1-\gamma} \end{pmatrix}.(11)✦ K1(γ,ν)=(0γ(1−ν)00).K_1(\gamma,\nu)= \begin{pmatrix}0&\sqrt{\gamma(1-\nu)}\\0&0\end{pmatrix}.(12)✦ K2(γ,ν)=(ν1−γ00ν).K_2(\gamma,\nu)= \begin{pmatrix}\sqrt\nu\sqrt{1-\gamma}&0\\0&\sqrt\nu\end{pmatrix}.(13)✦ K3(γ,ν)=(00γν0).K_3(\gamma,\nu)= \begin{pmatrix}0&0\\\sqrt{\gamma\nu}&0\end{pmatrix}.(14)✦ Let γ,ν∈[0,1]\gamma,\nu\in[0,1] and let pp be an optimal excitation population. For every nonnegative integer nn, the Holevo information satisfies (15), and the operational capacity satisfies (16): χ(𝒩γ,ν⊗n)=nln⁡2f(γ,ν,p).\chi(\mathcal N_{\gamma,\nu}^{\otimes n}) =\frac{n}{\ln2}f(\gamma,\nu,p).(15)✦ C(𝒩γ,ν)=f(γ,ν,p)ln⁡2.C(\mathcal N_{\gamma,\nu})=\frac{f(\gamma,\nu,p)}{\ln2}.(16)✦

In the paper, the result reads: for γ∈[0,1]\gamma\in[0,1], let AγA_\gamma have Kraus matrices K0=diag⁡(1,1−γ)K_0=\operatorname{diag}(1,\sqrt{1-\gamma}) and K1=γ|0⟩⟨1|K_1=\sqrt\gamma\,|0\rangle\langle1|; for every integer n≥1n\ge1, its Holevo information and capacity satisfy (17): χ(Aγ⊗n)=nC(Aγ)=nln⁡2maxp∈[0,1]{h((1−γ)p)−g(γ(1−γ)p2)}.\chi(A_\gamma^{\otimes n})=nC(A_\gamma) =\frac{n}{\ln2}\max_{p\in[0,1]} \left\{h((1-\gamma)p)-g(\gamma(1-\gamma)p^2)\right\}.(17)

What Is Checked

The Kraus identities are stated for real parameters, while the information identities concern tensor powers of one generalized amplitude-damping channel with γ,ν∈[0,1]\gamma,\nu\in[0,1], an optimal population p∈[0,1]p\in[0,1], finite quantum ensembles, and classical codes with vanishing average decoding error.