Consider calculating the probability of the outcome m alone of some composite system $AB$. \begin{align*} p_A(m) &= \sum_{n=0}^{d_B-1} p_{AB}(m,n)\,,\\ &= \sum_{n=0}^{d_B-1}(⟨α_m|⊗⟨β_n|)\rho_{AB}(|α_m⟩⊗|β_n⟩)\,. \end{align*}

I'm trying to show that this value is independent of the basis $\{|β_n\rangle\}\,,$ where $n = 0, 1, \cdots , d_B-1$ (so we can just replace it with the computational basis for example.

I know that I've to use the no signalling theorem for it, and will have to decompose the density matrix $\rho_{AB}$ into $\sum_i p_i |\psi_i⟩⟨\psi_i|$ but beyond that I haven't made much progress


1 Answer 1


Independence of the slightly more general expression $\sum_n\langle x\otimes \beta_n|\rho_{AB}|y\otimes \beta_n\rangle$ from the orthonormal basis $\{\beta_n\}_n$ boils down to the fact that the same is true for the resolution of the identity ${\bf1}=\sum_n|\beta_n\rangle\langle \beta_n|$ (=this identity holds for all orthonormal bases). For all $x,y,\rho_{AB}$ we compute: \begin{align*} \sum_n\langle x\otimes \beta_n|\rho_{AB}|y\otimes \beta_n\rangle&=\sum_n{\rm tr}\big(|y\otimes \beta_n\rangle\langle x\otimes \beta_n|\rho_{AB}\big)\\ &=\sum_n{\rm tr}\big((|y\rangle\langle x|\otimes |\beta_n\rangle\langle \beta_n|)\rho_{AB}\big)\\ &={\rm tr}\Big(\Big(|y\rangle\langle x|\otimes \sum_n|\beta_n\rangle\langle \beta_n|\Big)\rho_{AB}\Big)={\rm tr}\big((|y\rangle\langle x|\otimes {\bf1})\rho_{AB}\big) \end{align*} This shows that the expression $p_A(m)={\rm tr}((|\alpha_m\rangle\langle \alpha_m|\otimes {\bf1})\rho_{AB})$ is indeed independent of the chosen orthonormal basis on $B$, as claimed.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.