Bell nonlocality in quantum networks with unreliable sources: Loophole-free postelection via self-testing
Sadra Boreiri, Nicolas Brunner, Pavel Sekatski · 2025
Read Bell nonlocality in quantum networks with unreliable sources: Loophole-free postelection via self-testing by Sadra Boreiri, Nicolas Brunner, Pavel Sekatski (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-2501-14027.txt.
Introduction
Sadra Boreiri, Nicolas Brunner, and Pavel Sekatski tighten the rules for quantum network tests with flaky sources in this 2025 physics paper. In Bell tests, throwing away unclear events can open a loophole that fakes quantum effects. The authors derive a condition on the data that makes discarding those events safe.
They characterize when measurements fairly sample outcomes in a network. If all measurements qualify, keeping only clear events cannot fake source independence, so the detection loophole stays shut. Sometimes the data alone certify this fairness: saturating a known inequality self-tests the setup. They show the idea improves random number generation in a photonic test and a triangle network.
Catalog record
| CREATOR | Sadra Boreiri, Nicolas Brunner, Pavel Sekatski |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-14027
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Bell nonlocality in quantum networks with unreliable sources: Loophole-free postelection via self-testing Sadra Boreiri, Nicolas Brunner, Pavel Sekatski Abstract We discuss Bell nonlocality in quantum networks with unreliable sources. Our main result is a condition on the observed data which ensures that inconclusive events can be safely discarded, without introducing any loophole. More formally, we characterize the fair-sampling property for measurements in a network. When all measurements are fair-sampling, we show that the post-selection of conclusive outcomes does not compromise the assumption of source independence, hence avoiding the detection loophole. Furthermore, we show that in some cases, the fair-sampling property can in fact be guaranteed based only on observed data. To show this, we prove that saturation of the Finner inequality provides a self-test of the underlying quantum model. We illustrate the relevance of our results by demonstrating an improvement in device-independent randomness generation for a photonic Bell test with a probabilistic source and for the triangle network. I Introduction Bell nonlocality is now widely recognized as a central aspect of quantum theory, playing a vital role not only in deepening our understanding of the quantum world but also in paving the way for advanced applications in quantum information processing, leading to the device-independent paradigm review . In his seminal work bell1964einstein , Bell demonstrated that two distant observers performing local measurements on a shared entangled state can establish strong correlations that cannot be explained by any physical theory satisfying a natural notion of locality, respected in classical physics. Such Bell nonlocal quantum correlations have been experimentally demonstrated Hensen , and are the basis for several applications, most notably within the device-independent framework Acin2007 ; colbeck2009quantum ; pironio2010random . In recent years, the effect of quantum Bell nonlocality has been investigated in the context of quantum networks where several independent sources distribute quantum resources (e.g. entangled states) to the observers, see e.g. Branciard2010 ; Fritz_2012 ; Chaves2015 ; Gisin_2019 ; Renou_2019 ; boreiri2023towards ; Tavakoli_Review . The network setting allows for novel forms of Bell nonlocality, in particular in scenarios with fixed measurement settings (i.e. without requiring the observers to actively change their measurement setting) Fritz_2012 ; branciard2012bilocal . In this scenario, Bell nonlocality can be used to certify the combination of entangled states and joint entangled measurements Renou_2019 ; renou2022network ; Supic2022 ; Sekatski2023 , as well as for generating randomness (without the need for input randomness). So far, these questions have been mostly discussed in an idealized setting (e.g. considering deterministic sources of entanglement), and characterizing Bell nonlocality in more realistic networks remains challenging wolfe2019inflation ; krivachy_neural_2020 ; wolfe2021inflation ; boreiri2023noise . This is indeed central towards any future experiments or applications. In the present work, we discuss Bell nonlocality in networks with unreliable sources. This means that each source in the network, in each round of the Bell test, has a certain probability of failing, e.g. not sending out any particles. In case of such a failure, the observers connected to the source will report a failure (no-click) outcome. Dealing with such inconclusive outcomes in the context of Bell nonlocality is not trivial. In general, the naive strategy of simply discarding these inconclusive events and analyzing the nonlocality of the post-selected data is not valid. Indeed, this procedure can open the so-called detection loophole, where a local model can reproduce the (seemingly nonlocal) post-selected data by exploiting the inconclusive events pearle1970 ; review . This problem is related to the fact that Bell nonlocality is predicated on certain assumptions of causal independence. In the traditional Bell scenario, the choice of measurement setting must be independent of the source of entanglement, while in network nonlocality (where measurement settings are no longer necessary), one must ensure that the different sources are independent, in accordance with the network topology. From this perspective, the detection loophole arises when the post-selection breaks the causal independence relations. In this case, the post-selected data no longer comply with the network topology, hence invalidating the proof of Bell nonlocality. Nevertheless, here we prove that in the failing-source scenario, post-selection of inconclusive events is in fact harmless, and does not open any loophole. More precisely, we show that the correlations observed in this scenario self-test the underlying failing source model, i.e. the latter is the only quantum model compatible with the observed data. In this model, the occurrence of failure is decided at the source. Therefore post-selection does not compromise the assumption of source independence. After proving our general result, we illustrate its relevance with two examples. In particular, we apply our result to the standard photonic CHSH Bell test (with a probabilistic SPDC source of entanglement, and demonstrate its usefulness for increasing the rate of device-independent randomness generation. From a more technical perspective, the main contributions of our work are the following. First, we define the concept of fair-sampling measurements in the context of quantum networks, following the approach of berry2010fair ; orsucci2020post . In turn, we show that whenever all measurements in the network are fair-sampling, then post-selection is harmless. Finally, to prove our main result, we prove a rigidity (self-test) statement for all correlations saturating the Finner inequality finner1992generalization ; renou2019limits . Specifically, this rigidity property guarantees that all measurements are fair-sampling. Finally, we discuss the connection of our work with the effect of topologically robust network Bell nonlocality boreiri2024topologically and conclude with open questions. II Unreliable quantum networks and the fair sampling assumption Consider a network with N sources S_{i} and M parties A_{j} . The network can be represented by a bipartite graph where the connections between the sources and the parties are specified by the N\times M adjacency matrix E taking the values E_{ij}=1 (connection) E_{ij}=0 (no connection). Each party A_{j} performs a measurement on the systems received from the connected sources in order to produce an outcome a_{j} . In principle, before doing so it can also sample a measurement input x_{j} from a local source of randomness. However, any such network with inputs can be mapped into a “dressed” network without inputs by adding a source of randomness and an extra party announcing x_{j} (see the example in the left panel of Fig. 3 ). Thus, without loss of generality, we focus here on networks without inputs. The set of possible probability distributions of outputs {\rm P}(a_{1},a_{2},\dots,a_{M}) for a given network depends on the underlying physical theory governing the experiment. Classically (local models), a source distributes a random variable, which are read out by the measurements at the parties to produce outputs. In quantum theory, each source S_{i} sends a quantum system Q_{j}^{(i)} to the connected party A_{j} . All systems sent by a source S_{i} are denoted by the composite system \bm{Q}^{(i)}=\bigtimes_{j|E_{ij}=1}Q_{j}^{(i)} . The state prepared by the source S_{i} admits a purification \ket{\Psi^{(i)}}_{\bm{Q}^{(i)}E^{(i)}} , where E^{(i)} is a purifying system potentially controlled by an eavesdropper. In turn, the measurement performed by each party are given by a POVM \{M^{a_{j}}_{A_{j}}\} acting on the tensor product Hilbert space for all systems received by the party A_{j} . To discuss networks with unreliable elements (such as failing sources), we add to each measurement an inconclusive outcome, a_{j}=\varnothing , which corresponds to a failure (no-click) event reported by party A_{j}