Evaluation of Noise and Crosstalk in Neutral Atom Quantum Computers
Pranet Sharma, Yizhuo Tan, Konstantinos-Nikolaos Papadopoulos, Jakub Szefer · 2025
Read Evaluation of Noise and Crosstalk in Neutral Atom Quantum Computers by Pranet Sharma, Yizhuo Tan, Konstantinos-Nikolaos Papadopoulos, Jakub Szefer (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-2507-22140.txt.
Introduction
Pranet Sharma and co-authors measure noise and crosstalk in neutral atom quantum computers, in a 2025 paper. These machines are promising for simulating larger systems with time-varying energy rules. Such simulations serve work in optimization, physics, and molecular dynamics. Running several at once could save resources.
The team track how noise changes over time and test how spatial closeness affects simulation fidelity. Close concurrent runs raised crosstalk between them. They propose a moving target defense, which shifts the setup to reduce that interference, and test it. Results confirm the defense can support safer shared use of one machine.
Catalog record
| CREATOR | Pranet Sharma, Yizhuo Tan, Konstantinos-Nikolaos Papadopoulos, Jakub Szefer |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 16 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-22140
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Evaluation of Noise and Crosstalk in Neutral Atom Quantum Computers Thanks: † This work was supported in part by NSF grant 2332406 . Pranet Sharma, Yizhuo Tan, Konstantinos-Nikolaos Papadopoulos, Jakub Szefer Abstract This work explores and evaluates noise and crosstalk in neutral atom quantum computers. Neutral atom quantum computers are a promising platform for analog Hamiltonian simulations, which rely on a sequence of time-dependent Hamiltonians to model the dynamics of the larger system and are particularly useful for problems in optimization, physics, and molecular dynamics. However, the viability of running multiple simulations in a co-located or multi-tenant environment is limited by noise and crosstalk. This work conducts an analysis of how noise faced by simulations changes over time, and investigates the effects of spatial co-location on simulation fidelity. Findings of this work demonstrate that the close proximity of concurrent simulations can increase crosstalk between them. To mitigate this issue, a Moving Target Defense (MTD) strategy is proposed and evaluated. The results confirm that the MTD is a viable technique for enabling safe and reliable co-location of simulations on neutral atom quantum hardware. I Introduction The field of quantum computing has transitioned from a realm of theoretical promise to one of tangible, rapidly advancing hardware. Among the diverse modalities being pursued, neutral atom quantum computers have distinguished themselves as a compelling platform. Within these computers, atoms of alkali and alkaline earth metals such as rubidium, strontium, and cesium serve as qubits. These atoms are first confined in a magneto-optical trap, and are then loaded into optical tweezers [ 1 ] , which are used to move the atoms around for computation by exploiting the Rydberg blockade effect [ 2 ] . Due to the constant reconfiguration of qubits, the time cost of a quantum gate on current neutral atom quantum systems is high. However, this also makes them well-suited for the computational paradigm of analog Hamiltonian simulations (AHS), one of the first applications of quantum computing that has demonstrated quantum advantage when modeling problems in optimization, physics, and molecular dynamics [ 3 ] [ 4 ] [ 5 ] . To maximize the computational efficiency and throughput of neutral atom systems running these simulations, they can use multi-tenancy, or co-location. Here, the large number of available qubits on a single quantum “chip” are partitioned to run multiple, independent simulations in parallel. However, before such parallel execution can be reliably implemented, an engineering challenge must be overcome: the effects of system noise and inter-simulation crosstalk. Uncontrolled noise can affect the quantum states of a simulation, while crosstalk – unwanted quantum mechanical interactions between supposedly independent programs – can create spurious correlations, compromising the results. The impacts of crosstalk and the ways that it can be potentially weaponized in cloud-based quantum computers have been studied in the recent past [ 6 ] . However, not much work has been done toward understanding possible noise and crosstalk in neutral atom quantum computers, nor developing runtime protection against crosstalk. (a) Test on May 3, 2025. (b) Test on May 10, 2025. (c) Test on May 17, 2025. (d) Test on May 24, 2025. Fig. 1 : Temporal noise evaluation on the Aquila quantum computer over period of four weeks in May, 2025. The darker green colors represent parts of the array that performed more closely to the determined control. Uniform coloring – such as visible in Fig 1d – implies less variance in final qubit counts across multiple runs, while non-uniform coloring – such as visible in Fig 1a – implies more variance. II Simulation Framework and Control Setup II-A Analog Hamiltonian Simulations An AHS program relies on a sequence of time-dependent Hamiltonians governing the dynamics of the larger system. Neutral atom quantum computers execute AHS by mapping the time-dependent Hamiltonian evolution of their qubits (a quantum many-body Hamiltonian) to a real-life problem. Within this work, we used QuEra’s neutral atom backends. QuEra offers three degrees of freedom to control the quantum many-body Hamiltonian run on their backends: \begin{split}H(t)&=\sum_{k=1}^{N}H_{\text{drive},k}(t)\\ &\quad+\sum_{k=1}^{N}H_{\text{shift},k}(t)\\ &\quad+\sum_{j=1}^{N-1}\sum_{k=j+1}^{N}H_{\text{vdW},j,k}.\end{split} (1) where H_{\text{drive},k}(t) is the driving field, H_{\text{shift},k}(t) is the shifting field, and H_{\text{vdW},j,k} represents van der Waals interactions. The driving and shifting fields are externally applied to the simulation, while the van der Waals field is internal to the simulation and governs the interactions between each pair of qubits in the simulation [ 1 ] . The driving field is uniformly applied to every qubit in the register. Therefore, in our experiments, we are more interested in the shifting field, as it is applied directly to particular qubits as specified by the user. The impact of the shifting field creates an effect on the qubit known as “detuning” – shifting its frequency away from resonance – which changes the van der Waals interactions of that qubit, thereby inducing a form of crosstalk [ 7 ] . The shifting field can be further defined as H_{\text{shift},k}(t)=-\Delta_{\text{local}}(t)h_{k}\,n_{k}, (2) where \Delta_{\text{local}}(t) is the time-dependent magnitude of the frequency shift, and h_{k} is the atom-dependent pattern [ 1 ] . A carefully chosen \Delta_{\text{local}} value, therefore, can help simulate several real-world situations. On the other hand, an incorrectly or maliciously set shifting field can also be selected to induce the most disruption in nearby qubits – which could be abused in the case of multiple simulations co-located on one quantum computer in a multi-tenant setting. II-B Experimental Setup We used the same core simulation qs_{C} in every experiment. We accessed two backends in this experiment – Aquila, QuEra’s cloud-accessible neutral atom QPU, and the Amazon Web Services (AWS) Braket AHS simulator for QuEra [ 8 ] . Both backends were accessed via qBraid Lab [ 9 ] . For our experiments, we set up a three-qubit register forming an equilateral triangle with side lengths of 5.5 µm, selected as it is the starting atomic separation distance provided by QuEra for AHS [ 1 ] . For our experiments delivered to the simulator, we applied a shifting field to the top qubit in this register defined by \Delta_{\text{local}}=5\times 10^{7}\text{ rad/sec} , which was 20 times the frequency of the uniform driving field. In our result, therefore, we expected the top qubit would not be excited due to its detuning, while the other two were likely to become entangled due to their van der Waals interactions. Within AHS, results are quantified by the atom count of each qubit after the conclusion of the simulation. Therefore, with the introduction of our shifting field, we can expect extremely low counts for the detuned qubit and high, even counts for the other two. In contrast, without the introduction of a shifting field, we expect more evenly distributed counts. We did not include this specified shifting field when submitting to Aquila, as it currently does not support local detuning without specialized Braket Direct access; in future work, we aim to replicate this setup through that service [ 10 ] . We began by establishing our control: expected qubit counts for qs_{C} , determined by averaging individual qubit counts over multiple runs on the simulator. In each experiment, we defined relative fidelity as the averaged percent difference in final qubit counts for each qubit in our experimentally run simulations from the expectation. The closer