Lean, informalized

Nielsen & Chuang · Chapter 12, Quantum information theory · p. 535

Exercise 12.3 - What a measurement on nn qubits reveals about the message they encode is at most nn bits

Each qubit carries at most one bit.

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 Exercise

Alice picks a message xx from a finite set, each xx with probability pxp_x, and encodes it in nn qubits by preparing them in a state ρx\rho_x. She sends the qubits to Bob, who performs a measurement on them and obtains one of rr outcomes yy. Message and outcome are then correlated random variables XX and YY, and their mutual information H(X:Y)H(X:Y) is what Bob learns about the message. The exercise is to show that he never learns more than nn bits, whatever states Alice prepares and whatever measurement he makes.✦ It is a consequence of the Holevo bound.

A Quantum System

A quantum system SS is a physical system whose state space is a finite dimensional complex Hilbert space ℋS\mathcal{H}_S.✦ The qubit QQ is the two level system, the system whose state space is ℂ2\mathbb{C}^2.✦ For finitely many systems SiS_i, the composite system ⨂iSi\bigotimes_i S_i is the system whose state space is the tensor product ⨂iℋSi\bigotimes_i \mathcal{H}_{S_i}✦, and the tensor power S⊗nS^{\otimes n} of a system SS is the composite of nn copies of SS.✦ Alice’s nn qubits form the system Q⊗nQ^{\otimes n}, with state space ℋQ⊗n=(ℂ2)⊗n.\mathcal{H}_{Q^{\otimes n}} = (\mathbb{C}^2)^{\otimes n}.(1)

A state ρ\rho of SS is a positive operator of unit trace on ℋS\mathcal{H}_S.✦ Each ρx\rho_x is a state of Q⊗nQ^{\otimes n}.

A measurement MM on SS with a finite set of outcomes is a family of operators MyM_y on ℋS\mathcal{H}_S, one for each outcome yy, satisfying the completeness relation ∑yMy†My=I,\sum_y M_y^\dagger M_y = I,(2)✦ where II is the identity on ℋS\mathcal{H}_S. The positive operator valued measure (POVM) of MM is the family of operators Ey=My†MyE_y = M_y^\dagger M_y on ℋS\mathcal{H}_S, the effects of the outcomes.✦ The probability of the outcome yy when MM is performed on a state ρ\rho of SS is given by the Born rule, p(y∣ρ)=Tr⁡(MyρMy†)=Tr⁡(Eyρ).p(y \mid \rho) = \operatorname{Tr}(M_y \rho M_y^\dagger) = \operatorname{Tr}(E_y \rho).(3)✦

The Claim

The message XX and the outcome YY have the joint distribution p(x,y)=pxTr⁡(Eyρx).p(x, y) = p_x \operatorname{Tr}(E_y \rho_x).(4) H(X:Y)H(X:Y) is the mutual information of this distribution, with logarithms to base ee; nn bits are then nlog⁡2n \log 2.

Theorem 1. Let pxp_x be a probability distribution on a finite set of messages, let ρx\rho_x be a state of nn qubits for each message xx, and let MM be a measurement on nn qubits with r≥1r \geq 1 outcomes. Then the message XX and the outcome YY, with the joint distribution (4), satisfy H(X:Y)≤nlog⁡2.H(X:Y) \leq n \log 2.(5)✦

What Is Checked

The claim is stated over the system Q⊗nQ^{\otimes n} of nn qubits, the tensor power of the qubit, over states of that system, and over a measurement on it.

The library proves that the effects EyE_y of a measurement are positive, so that they form a POVM. It proves also that a composite of finitely many systems has a finite dimensional state space, since the products of basis vectors of the factors form a basis of the tensor product; Q⊗nQ^{\otimes n} is thus a quantum system.