OctoInformBeta

OpenAI preprint, September 2026 · 5. The ensemble bound and its attainment · Section 5

The ensemble bound and phase-product attainment: Arbitrary block signals and a uniform ensemble of phase products for a fixed generalized amplitude-damping channel

Optimal phase products attain the ensemble bound.

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 repeated uses of a generalized amplitude-damping channel, a signal may be an arbitrary density operator on the whole block. The comparison with product signals concerns the Holevo information of a finite ensemble after transmission through the same channel at every site.

Let nn be a nonnegative integer. The computational basis ℬn\mathcal B_n consists of bit strings Fin⁡(n)→Fin⁡(2)\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 with rows and columns indexed by ℬn\mathcal B_n.✦ A density operator PP on a finite-dimensional space is a positive semidefinite complex matrix of trace one.✦

For damping γ∈[0,1]\gamma\in[0,1] and stationary excited-state population ν∈[0,1]\nu\in[0,1], the single-qubit Kraus operators are the matrices in (1): K0(1)=1−ν(1001−γ),K1(1)=γ(1−ν)(0100),K2(1)=ν(1−γ001),K3(1)=γν(0010).\begin{aligned} K^{(1)}_0&=\sqrt{1-\nu} \begin{pmatrix}1&0\\0&\sqrt{1-\gamma}\end{pmatrix},& K^{(1)}_1&=\sqrt{\gamma(1-\nu)} \begin{pmatrix}0&1\\0&0\end{pmatrix},\\ K^{(1)}_2&=\sqrt{\nu} \begin{pmatrix}\sqrt{1-\gamma}&0\\0&1\end{pmatrix},& K^{(1)}_3&=\sqrt{\gamma\nu} \begin{pmatrix}0&0\\1&0\end{pmatrix}. \end{aligned}(1)✦ The tensor product Kraus operator KrK_r with label r∈Fin⁡(n)→Fin⁡(4)r\in\operatorname{Fin}(n)\to\operatorname{Fin}(4) has the entries in (2), for i,j∈ℬni,j\in\mathcal B_n: Kr(i,j)=∏k=0n−1Krk(1)(ik,jk).K_r(i,j)=\prod_{k=0}^{n-1}K^{(1)}_{r_k}(i_k,j_k).(2)✦ The memoryless channel 𝒩γ,ν⊗n\mathcal N_{\gamma,\nu}^{\otimes n} is the nn-fold tensor power of the fixed single-qubit channel and acts on an nn-qubit operator PP by (3); the star denotes the adjoint: 𝒩γ,ν⊗n(P)=∑rKrPKr*.\mathcal N_{\gamma,\nu}^{\otimes n}(P)=\sum_r K_r P K_r^*.(3)✦

For a finite complex matrix PP, the von Neumann entropy S(P)S(P) is the real part of the trace of the spectral functional calculus with the scalar function x↦−xln⁡xx\mapsto-x\ln x.✦ For density operators it takes the form in (4), where SS and the binary entropy h(q)h(q) for q∈[0,1]q\in[0,1] are measured in nats and 0ln⁡0=00\ln0=0. S(P)=−Tr⁡(Pln⁡P),h(q)=−qln⁡q−(1−q)ln⁡(1−q).S(P)=-\operatorname{Tr}(P\ln P),\qquad h(q)=-q\ln q-(1-q)\ln(1-q).(4)

For a nonnegative integer mm, a finite quantum ensemble E={wa,ρa}a=0m−1E=\{w_a,\rho_a\}_{a=0}^{m-1} on nn qubits consists of density operators ρa\rho_a and nonnegative weights waw_a with ∑awa=1\sum_a w_a=1.✦ The output Holevo information χγ,ν(E)\chi_{\gamma,\nu}(E) of a finite quantum ensemble EE on nn qubits is the quantity in bits defined by (5): χγ,ν(E)=1ln⁡2[S(∑awa𝒩γ,ν⊗n(ρa))−∑awaS(𝒩γ,ν⊗n(ρa))].\chi_{\gamma,\nu}(E)=\frac{1}{\ln2}\left[ S\!\left(\sum_a w_a\mathcal N_{\gamma,\nu}^{\otimes n}(\rho_a)\right) -\sum_a w_a S\!\left(\mathcal N_{\gamma,\nu}^{\otimes n}(\rho_a)\right) \right].(5)✦

The local excitation population pj(P)p_j(P) of an nn-qubit operator PP at site j∈Fin⁡(n)j\in\operatorname{Fin}(n) is the real diagonal sum in (6): pj(P)=∑i∈ℬnij=1Re⁡P(i,i).p_j(P)=\sum_{\substack{i\in\mathcal B_n\\i_j=1}}\operatorname{Re}P(i,i).(6)✦

The Phase Ensemble and Scalar Objective

For input excitation population p∈[0,1]p\in[0,1], phase label σ∈Fin⁡(2)\sigma\in\operatorname{Fin}(2) and computational bit b∈Fin⁡(2)b\in\operatorname{Fin}(2), the binary phase signal amplitude ap(σ,b)a_p(\sigma,b) is the complex number specified by (7): ap(σ,b)={1−p,b=0,(−1)σp,b=1.a_p(\sigma,b)= \begin{cases} \sqrt{1-p},&b=0,\\ (-1)^\sigma\sqrt p,&b=1. \end{cases}(7)✦ For a phase string s∈ℬns\in\mathcal B_n, the product signal amplitude Ap,sA_{p,s} is a complex vector on ℬn\mathcal B_n, and its product signal state ρp,s\rho_{p,s} is the operator with entries in (8): Ap,s(i)=∏k=0n−1ap(sk,ik),ρp,s(i,j)=Ap,s(i)Ap,s(j)¯.A_{p,s}(i)=\prod_{k=0}^{n-1}a_p(s_k,i_k),\qquad \rho_{p,s}(i,j)=A_{p,s}(i)\overline{A_{p,s}(j)}.(8) The uniform phase ensemble En,pphE^{\mathrm{ph}}_{n,p} is the ensemble of product signal states ρp,s\rho_{p,s} indexed by s∈ℬns\in\mathcal B_n, with weight 2−n2^{-n} for each state.✦ The phase ensemble value Φγ,ν(p;n)\Phi_{\gamma,\nu}(p;n) of the uniform phase ensemble En,pphE^{\mathrm{ph}}_{n,p} is the entropy difference in bits defined in (9): Φγ,ν(p;n)=1ln⁡2[S(∑s∈ℬn2−n𝒩γ,ν⊗n(ρp,s))−∑s∈ℬn2−nS(𝒩γ,ν⊗n(ρp,s))].\Phi_{\gamma,\nu}(p;n)=\frac{1}{\ln2}\left[ S\!\left(\sum_{s\in\mathcal B_n}2^{-n} \mathcal N_{\gamma,\nu}^{\otimes n}(\rho_{p,s})\right) -\sum_{s\in\mathcal B_n}2^{-n} S\!\left(\mathcal N_{\gamma,\nu}^{\otimes n}(\rho_{p,s})\right) \right].(9)✦

The qubit entropy gg as a function of the determinant uu is defined by (10): g(u)=h(1+1−4u2).g(u)=h\!\left(\frac{1+\sqrt{1-4u}}{2}\right).(10)✦ For input excitation population p∈[0,1]p\in[0,1], the phase pair consists of the two binary phase signals with labels σ=0\sigma=0 and σ=1\sigma=1; the output determinant vv of either signal and the phase pair Holevo objective ff in nats are the functions in (11): 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}(11) An optimal excitation population p*p_* is a number in [0,1][0,1] satisfying (12): f(γ,ν,q)≤f(γ,ν,p*)(q∈[0,1]).f(\gamma,\nu,q)\leq f(\gamma,\nu,p_*)\qquad(q\in[0,1]).(12)✦

The Claim

Theorem 1. Let nn be a nonnegative integer. For real parameters γ,ν\gamma,\nu and population p∈[0,1]p\in[0,1], the phase ensemble identity is (13): χγ,ν(En,pph)=Φγ,ν(p;n).\chi_{\gamma,\nu}(E^{\mathrm{ph}}_{n,p})=\Phi_{\gamma,\nu}(p;n).(13)✦

Suppose γ,ν∈[0,1]\gamma,\nu\in[0,1]. For any nn-qubit density operator PP, the output entropy bound is (14): S(𝒩γ,ν⊗n(P))≤∑j=0n−1h((1−γ)pj(P)+γν).S\!\left(\mathcal N_{\gamma,\nu}^{\otimes n}(P)\right) \leq\sum_{j=0}^{n-1}h((1-\gamma)p_j(P)+\gamma\nu).(14)✦ For any finite quantum ensemble E={wa,ρa}E=\{w_a,\rho_a\} on nn qubits and any optimal excitation population p*p_*, the ensemble Holevo numerator obeys (15): S(∑awa𝒩γ,ν⊗n(ρa))−∑awaS(𝒩γ,ν⊗n(ρa))≤nf(γ,ν,p*).S\!\left(\sum_a w_a\mathcal N_{\gamma,\nu}^{\otimes n}(\rho_a)\right) -\sum_a w_a S\!\left(\mathcal N_{\gamma,\nu}^{\otimes n}(\rho_a)\right) \leq n f(\gamma,\nu,p_*).(15)✦ For every population p∈[0,1]p\in[0,1], the phase value satisfies (16): Φγ,ν(p;n)=nln⁡2f(γ,ν,p).\Phi_{\gamma,\nu}(p;n)=\frac{n}{\ln2}f(\gamma,\nu,p).(16)✦

What Is Checked

The bounds (15) and (14) are stated over arbitrary finite ensembles of density operators and arbitrary block density operators, respectively, while (13) and (16) concern the uniform ensemble of phase product states. For p∈[0,1]p\in[0,1], the phase product signals are density operators, and their uniform weights are nonnegative and sum to one.