Factorizability of optimal quantum sequence discrimination under maximum-confidence measurements
Donghoon Ha, Jeong San Kim · 2025
Read Factorizability of optimal quantum sequence discrimination under maximum-confidence measurements by Donghoon Ha, Jeong San Kim (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-20311.txt.
Introduction
Donghoon Ha and Jeong San Kim ask if best tests of quantum state chains can split by step. The tests use max-confidence rules, where the aim is to name a state with top trust. Their answer is yes, in full. The best test of a whole chain is just the best test at each step, done one by one.
The top trust for a chain comes from the top trust at each state in it. No joint test can beat that split form. The pair also give a full if-and-only-if test for best state tests under max-confidence rules. That test rounds out the proof and guides real use.
Catalog record
| CREATOR | Donghoon Ha, Jeong San Kim |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-20311
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Factorizability of optimal quantum sequence discrimination under maximum-confidence measurements Donghoon Ha, Jeong San Kim Abstract We consider the discrimination of quantum sequences under maximum-confidence measurements and show that the optimal discrimination of a quantum sequence ensemble can always be factorized into that of each individual ensemble. In other words, the optimal quantum sequence discrimination under maximum-confidence measurements can be achieved just by performing a maximum-confidence discrimination independently at each step of the quantum sequence. We also show that the maximum confidence of identifying a quantum sequence is to achieve the maximum confidence of identifying each state comprising the quantum sequence. We further provide a necessary and sufficient condition for the optimal quantum state discrimination under maximum-confidence measurements. I Introduction Quantum state discrimination is one of the fundamental concepts used in various quantum information processing tasks chef2000 ; berg2007 ; barn20091 ; bae2015 . In general, we can always discriminate orthogonal quantum states using an appropriate measurement. On the other hand, no measurement can perfectly discriminate nonorthogonal quantum states. For these reasons, various state discrimination strategies have been developed to optimally discriminate nonorthogonal quantum states, such as minimum-error discrimination, unambiguous discrimination and maximum-confidence discrimination hels1969 ; ivan1987 ; diek1988 ; pere1988 ; crok2006 . When a discrimination process involves multiple quantum state ensembles, the situation becomes significantly more intricate, as it requires identifying not just a single quantum state but a sequence of states where each state is independently prepared from a quantum state ensemble. Quantum sequence discrimination naturally appears in various quantum information tasks, including multiple-copy state discrimination and quantum change-point detection higg2011 ; sent2016 ; sent2017 . Thus, a deeper understanding of quantum sequence discrimination provides valuable insights into the general principles of quantum information processing. The optimal discrimination of a quantum sequence ensemble can be factorized into that of each individual ensemble. In other words, the optimal discrimination of quantum sequences can be realized just by performing a quantum state discrimination independently at each step of the quantum sequence. This factorizability always holds in minimum-error discrimination and optimal unambiguous discrimination gupt20241 ; gupt20242 . Recently, a factorizability condition was established in the optimal discrimination of multi-party quantum sequences under local operations and classical communication ha2025 . The factorizability of quantum sequence discrimination characterizes the fundamental limits of quantum information processing tasks such as quantum cryptography, quantum teleportation and quantum data hiding benn1986 ; benn1993 ; divi2002 . When a quantum information processing task based on quantum state discrimination is independently repeated multiple times, the factorizability inherent in the underlying discrimination strategy implies that quantum memory or collective measurements offer no advantage. As there exist quantum information processing tasks based on maximum-confidence discrimination benn1992 ; neve2012 , it is a natural question to ask whether factorizability still holds in the optimal quantum sequence discrimination under maximum-confidence measurements. Here, we provide an answer to the question by showing that the optimal discrimination of a quantum sequence ensemble under maximum-confidence measurements can always be factorized into that of each individual ensemble. In other words, the optimal quantum sequence discrimination under maximum-confidence measurements can be achieved just by performing a maximum-confidence discrimination independently at each step of the quantum sequence. We also show that the maximum confidence of identifying a quantum sequence is to achieve the maximum confidence of identifying each state in the quantum sequence. Moreover, we establish a necessary and sufficient condition for the optimal quantum state discrimination under maximum-confidence measurements. This paper is organized as follows. In Sec. II.1 , we review the definitions and properties of the confidence of measurement in quantum state discrimination. In Sec. II.2 , we provide a necessary and sufficient condition for the optimal quantum state discrimination under maximum-confidence measurements. In Sec. III.1 , we provide some properties of the confidence of measurement in quantum sequence discrimination, and show that the maximum confidence of identifying a quantum sequence is to achieve the maximum confidence of identifying each state comprising the quantum sequence. In Sec. III.2 , we show that the optimal discrimination of a quantum sequence ensemble under maximum-confidence measurements can always be factorized into that of each individual ensemble. In Sec. IV , we summarize our results, discuss a useful application, and suggest a direction for future research. II Maximum-confidence discrimination of quantum states For a finite-dimensional complex Hilbert \mathcal{H} , let \mathbb{H} be the set of all Hermitian operators acting on \mathcal{H} . We also denote by \mathbb{H}_{+} the set of all positive-semidefinite operators in \mathbb{H} , that is, \mathbb{H}_{+}=\{E\in\mathbb{H}\,|\,\bra{v}E\ket{v}\geqslant 0~\mbox{for all}~\ket{v}\in\mathcal{H}\}. (1) A quantum state is described by a density operator \rho , that is, \rho\in\mathbb{H}_{+} with \Tr\rho=1 . A measurement is expressed by a positive operator-valued measure \{M_{i}\}_{i} , that is, \{M_{i}\}_{i}\subseteq\mathbb{H}_{+} with \sum_{i}M_{i}=\mathbbm{1} where \mathbbm{1} is the identity operator in \mathbb{H} . For a quantum state \rho , performing a measurement \{M_{i}\}_{i} gives the measurement outcome corresponding to M_{i} with probability \Tr(\rho M_{i}) . II.1 Quantum state discrimination and confidence of measurement Let us consider the situation of discriminating n quantum states \rho_{1},\ldots,\rho_{n} where the state \rho_{i} is prepared with the nonzero probability \eta_{i} for each i=1,\ldots,n . We denote this situation as an ensemble \mathcal{E}=\{\eta_{i},\rho_{i}\}_{i\in\mathbb{N}_{n}} (2) where \mathbb{N}_{n} is the set of all integers from 1 to n , that is, \mathbb{N}_{n}=\{1,\ldots,n\}. (3) The state of a quantum system prepared from \mathcal{E} is denoted by \rho_{0} , that is, \rho_{0}=\sum_{i\in\mathbb{N}_{n}}\eta_{i}\rho_{i}. (4) To guess the prepared state from the ensemble \mathcal{E} in Eq. ( 2 ), we apply the decision rule using a measurement \mathcal{M}=\{M_{?}\}\cup\{M_{i}\}_{i\in\mathbb{N}_{n}}. (5) For each i\in\mathbb{N}_{n} , obtaining the measurement outcome corresponding to M_{i} leads us to guess the prepared state as \rho_{i} . On the other hand, the measurement outcome corresponding to M_{?} is inconclusive: “I don’t know what state is prepared.” In this case, the average probabilities of correctly guessing the prepared state is \sum_{i\in\mathbb{N}_{n}}\eta_{i}\Tr(\rho_{i}M_{i}). (6) Now, we introduce the confidence of a measurement in quantum state discrimination. Definition 1 . For an ensemble \mathcal{E}=\{\eta_{i},\rho_{i}\}_{i\in\mathbb{N}_{n}} , a measurement \mathcal{M}=\{M_{?}\}\cup\{M_{i}\}_{i\in\mathbb{N}_{n}} and x\in\mathbb{N}_{n} , the confidence of \mathcal{M} to identify \rho_{x} is the conditional probability \Pr(\rho_{x}|M_{x})=\frac{\eta_{x}\Tr(\rho_{x}M_{x})}{\Tr(\rho_{0}M_{x})}, (7) which is well defined only when \Tr(\rho_{0}M_{x})>0. (8) The maximum confidence of measurements to identify \rho_{x} is \mathcal{C}_{x}(\mathcal{E})=\max_{\begin{subarray}{c}\mathcal{M}~\mathsf{with}\\ \Tr(\rho_{0}M_{x})>0\end{subarray}}\Pr(\rho_{x}|M_{x}), (9) where the maximum is taken over all possible measurements \mathcal{M}=\{M_{?}\}\cup\{M_{i}\}_{i\in\mathbb{N}_{n}} with Eq. ( 8 ). Here, we note that the confidence \Pr(\rho_{x}|M_{x}) in Eq. ( 7 ) represents the probability that the prepared state is \rho_{x} given that the measurement outcome corresponding to M_{x} is obtained. We also note that the maximum confidence \mathcal{C}_{x}(\mathcal{E}) in Eq. ( 9 ) is known to be the largest eigenvalue of \sqrt{\rho_{0}}^{-1}\eta_{x}\rho_{x}\sqrt{\rho_{0}}^{-1} where \sqrt{\rho_{0}} is the positive square root of \rho_{0} and \sqrt{\rho_{0}}^{-1} is the inverse of \sqrt{\rho_{0}} on its support crok2006 ; supp . From this fact, we can easily see that \displaystyle\mathcal{C}_{x}(\mathcal{E}) \displaystyle= \displaystyle\min\{q\in\mathbb{R}\,|\,q\mathbbm{1}-\sqrt{\rho_{0}}^{-1}\eta_{x}\rho_{x}\sqrt{\rho_{0}}^{-1}\in\mathbb{H}_{+}\} (10) \displaystyle= \displaystyle\min\{q\in\mathbb{R}\,|\,q\rho_{0}-\eta_{x}\rho_{x}\in\mathbb{H}_{+}\} where \mathbb{R} is the set of all real numbers. For an ensemble \mathcal{E}=\{\eta_{i},\rho_{i}\}_{i\in\mathbb{N}_{n}} , the following lemma shows that the maximum confidence \mathcal{C}_{i}(\mathcal{E}) is bounded below by the nonzero probability \eta_{i} for all i\in\mathbb{N}_{n} . Lemma 1 . For an ensemble \mathcal{E}=\{\eta_{i},\rho_{i}\}_{i\in\mathbb{N}_{n}} and x\in\mathbb{N}_{n} , we have \mathcal{C}_{x}(\mathcal{E})=\max_{\begin{subarray}{c}E\in\mathbb{H}_{+}\\ \Tr(\rho_{0}E)=1\end{subarray}}\eta_{x}\Tr(\rho_{x}E)\geqslant\eta_{x}, (11) where \rho_{0} is defined in Eq. ( 4 ) and the maximum is taken