Quantum Physics (quant-ph)

Abstract: Quantum batteries are predicted to have the potential to outperform their classical counterparts and are therefore an important element in the development of quantum technologies. In this work we simulate the charging process and work extraction of many-body quantum batteries on noisy-intermediate scale quantum (NISQ) devices, and devise the Variational Quantum Ergotropy (VQErgo) algorithm which finds the optimal unitary operation that maximises work extraction from the battery. We test VQErgo by calculating the ergotropy of a quantum battery undergoing transverse field Ising dynamics. We investigate the battery for different system sizes and charging times and analyze the minimum required circuit depth of the variational optimization using both ideal and noisy simulators. Finally, we optimize part of the VQErgo algorithm and calculate the ergotropy on one of IBM's quantum devices.

Abstract: We evaluated the accuracy limit for estimating gravitational potential using optical lattice clocks by utilizing the quantum Cram\'{e}r--Rao bound. We then compared the results for single-layer and multilayer optical lattice clocks. The results indicate that the lower bound of variance of the estimator of gravitational potential using finite-size optical lattice clocks diverges and recovers repeatedly as a function of time. Namely, the accuracy of the gravitational potential estimation is not a monotonic function of time owing to the effect of gravitational dephasing in finite-size optical lattice clock. Further, this effect creates an estimation accuracy limit when attempting to avoid the divergence of the lower bound. When the number of layers in the optical lattice clock is sufficiently large, the limit is independent of the optical lattice clock details. The time required to reach this limit is calculated to be approximately 33 hours for a three-dimensional optical lattice clock consisting of one million cadmium atoms due to Earth's gravity, and approximately the same for other atoms.

Abstract: Boolean satisfiability (SAT) solving is a fundamental problem in computer science. Finding efficient algorithms for SAT solving has broad implications in many areas of computer science and beyond. Quantum SAT solvers have been proposed in the literature based on Grover's algorithm. Although existing quantum SAT solvers can consider all possible inputs at once, they evaluate each clause in the formula one by one sequentially, making the time complexity O(m) -- linear to the number of clauses m -- per Grover iteration. In this work, we develop a parallel quantum SAT solver, which reduces the time complexity in each iteration from linear time O(m) to constant time O(1) by utilising extra entangled qubits. To further improve the scalability of our solution in case of extremely large problems, we develop a distributed version of the proposed parallel SAT solver based on quantum teleportation such that the total qubits required are shared and distributed among a set of quantum computers (nodes), and the quantum SAT solving is accomplished collaboratively by all the nodes. We have proved the correctness of our approaches and demonstrated them in simulations.

Abstract: The conventional formulation of quantum sensing is based on the assumption that the probe is reset to its initial state after each measurement. In a very distinct approach, one can also pursue a sequential measurement scheme in which time-consuming resetting is avoided. In this situation, every measurement outcome effectively comes from a different probe, yet correlated with other data samples. Finding a proper description for the precision of sequential measurement sensing is very challenging as it requires the analysis of long sequences with exponentially large outcomes. Here, we develop a recursive formula and an efficient Monte-Carlo approach to calculate the Fisher information, as a figure of merit for sensing precision, for arbitrary lengths of sequential measurements. Our results show that Fisher information initially scales non-linearly with the number of measurements and then asymptotically saturates to linear scaling. Such transition, which fundamentally constrains the extractable information about the parameter of interest, is directly linked to the finite memory of the probe when undergoes multiple sequential measurements. Based on these, we establish a figure of merit to determine the optimal measurement sequence length and exemplify our results in three different physical systems.

Abstract: Quantum computers are becoming practical for computing numerous applications. However, simulating quantum computing on classical computers is still demanding yet useful because current quantum computers are limited because of computer resources, hardware limits, instability, and noises. Improving quantum computing simulation performance in classical computers will contribute to the development of quantum computers and their algorithms. Quantum computing simulations on classical computers require long performance times, especially for quantum circuits with a large number of qubits or when simulating a large number of shots for noise simulations or circuits with intermediate measures. Graphical processing units (GPU) are suitable to accelerate quantum computer simulations by exploiting their computational power and high bandwidth memory and they have a large advantage in simulating relatively larger qubits circuits. However, GPUs are inefficient at simulating multi-shots runs with noises because the randomness prevents highly parallelization. In addition, GPUs have a disadvantage in simulating circuits with a small number of qubits because of the large overheads in GPU kernel execution. In this paper, we introduce optimization techniques for multi-shot simulations on GPUs. We gather multiple shots of simulations into a single GPU kernel execution to reduce overheads by scheduling randomness caused by noises. In addition, we introduce shot-branching that reduces calculations and memory usage for multi-shot simulations. By using these techniques, we speed up x10 from previous implementations.

Abstract: The set of valid covariance matrices of a continuous-variable quantum system with a finite number of degrees of freedom is a strict subset of the set of real positive-definite matrices due to Heisenberg's uncertainty principle. This has the implication that, in general, not every orthogonal transform of a diagonal quantum covariance matrix produces a valid quantum covariance matrix. A natural question thus arises, to find the set of quantum covariance matrices consistent with a given eigenspectrum. For the special class of pure Gaussian states the set of quantum covariance matrices with a given eigenspectrum consists of a single orbit of the action of the orthogonal symplectic group. The eigenspectra of the covariance matrices of this class of states are composed of pairs that each multiply to one. We describe a larger non-trivial class of eigenspectra with the property that the set of quantum covariance matrices corresponding to any eigenspectrum in this class are related by orthogonal symplectic transformations. Further, we show that all non-degenerate eigenspectra with this property must belong to this class, and that the set of such eigenspectra coincides with the class of non-degenerate eigenspectra that identify the physically relevant thermal and squeezing parameters of a Gaussian state.

Abstract: Quantum mechanics has withstood every experimental test thus far. However, it relies on ad-hoc postulates which require experimental verification. Over the past decade there has been a great deal of research testing these postulates, with numerous tests of Born's rule for determining probabilities and the complex nature of the Hilbert space being carried out. Although these tests are yet to reveal any significant deviation from textbook quantum theory, it remains important to conduct such tests in different configurations and using different quantum states. Here we perform the first such test using coherent states of light in a three-arm interferometer combined with homodyne detection. Our proposed configuration requires additional assumptions, but allows us to use quantum states which exist in a larger Hilbert space compared to previous tests. For testing Born's rule, we find that the third order interference is bounded to be $\kappa$ = 0.002 $\pm$ 0.004 and for testing whether quantum mechanics is complex or not we find a Peres parameter of F = 1.0000 $\pm$ 0.0003 (F = 1 corresponds to the expected complex quantum mechanics). We also use our experiment to test Glauber's theory of optical coherence.

Abstract: The spinorial degrees of freedom of two spacelike separated Dirac particles are considered and a collection of nine locally Lorentz covariant bitensors is constructed. Four of these bitensors have been previously described in [Phys. Rev. A 105, 032402 (2022), arXiv:2103.07784]. The collection of bitensors has the property that all nine bitensors are simultaneously zero if and only if the state of the two particles is a product state. Thus this collection of bitensors indicates any type of spinor entanglement between two spacelike separated Dirac particles.

Abstract: We examine the relationship between the second law of thermodynamics and the energy eigenstates of quantum many-body systems that undergo cyclic unitary evolution. Using a numerically optimized control protocol, we analyze how the work extractability is affected by the integrability of the system. Our findings reveal that, in nonintegrable systems the number of work-extractable energy eigenstates converges to zero, even when the local control operations are optimized. In contrast, in integrable systems, there are exponentially many eigenstates from which positive work can be extracted, regardless of the locality of the control operations. We numerically demonstrate that such a strikingly different behavior can be attributed to the number of athermal energy eigenstates. Our results provide insights into the foundations of the second law of thermodynamics in isolated quantum many-body systems, which are expected to contribute to the development of quantum many-body heat engines.

Abstract: Nearest-neighbour clustering is a powerful set of heuristic algorithms that find natural application in the decoding of signals transmitted using the $M$-Quadrature Amplitude Modulation (M-QAM) protocol. Lloyd et al. proposed a quantum version of the algorithm that promised an exponential speed-up. We analyse the performance of this algorithm by simulating the use of a hybrid quantum-classical implementation of it upon 16-QAM and experimental 64-QAM data. We then benchmark the implementation against the classical k-means clustering algorithm. The choice of quantum encoding of the classical data plays a significant role in the performance, as it would for the hybrid quantum-classical implementation of any quantum machine learning algorithm. In this work, we use the popular angle embedding method for data embedding and the SWAP test for overlap estimation. The algorithm is emulated in software using Qiskit and tested on simulated and real-world experimental data. The discrepancy in accuracy from the perspective of the induced metric of the angle embedding method is discussed, and a thorough analysis regarding the angle embedding method in the context of distance estimation is provided. We detail an experimental optic fibre setup as well, from which we collect 64-QAM data. This is the dataset upon which the algorithms are benchmarked. Finally, some promising current and future directions for further research are discussed.

Abstract: A driven quantum system can experience Landau-Zener-Stueckelberg-Majorana (LZSM) transitions between its states, when the respective energy levels quasi-cross. If this quasicrossing is passed repeatedly under periodic driving, the trajectories can interfere either constructively or destructively. In the latter case, known as coherent destruction of tunneling, the transition between the energy states is suppressed. Even for a double-passage case, the accumulated phase difference (also referred to as the Stueckelberg phase) can lead to destructive interference, resulting in no transition. In this paper we discuss a similar process for a single-passage dynamics. We study the LZSM single-passage problem starting from a superposition state. The phase difference of this initial state results in interference. When this is destructive, resulting in a zero transition probability, such situation can be called single-passage coherent destruction of tunneling. When the phase is chosen so that the occupation probabilities do not change after the transition, this can be called occupation conservation and this is analogous to the problem of transitionless driving. We demonstrate how varying the system parameters (driving velocity, initial phase, initial detuning) can be used for quantum control.

Abstract: In this paper we explore the use of quantum machine learning (QML) applied to credit scoring for small and medium-sized enterprises (SME). A quantum/classical hybrid approach has been used with several models, activation functions, epochs and other parameters. Results are shown from the best model, using two quantum classifiers and a classical neural network, applied to data for companies in Singapore. We observe significantly more efficient training for the quantum models over the classical models with the quantum model trained for 350 epochs compared to 3500 epochs for comparable prediction performance. Surprisingly, a degradation in the accuracy was observed as the number of qubits was increased beyond 12 qubits and also with the addition of extra classifier blocks in the quantum model. Practical issues for executing on simulators and real quantum computers are also explored. Overall, we see great promise in this first in-depth exploration of the use of hybrid QML in credit scoring.

Abstract: By considering the local quantum uncertainty (LQU) as a measure of quantum correlations, the thermal evolution of a two-qubit-superconducting system is investigated. We show that the thermal LQU can be increased by manipulating the Hamiltonian parameters such as the mutual coupling and Josephson energies, however, it undergoes sudden transitions at specific temperatures. Notably, our theoretical results are in good agreement with experimental data for thermal entanglement. Furthermore, a detailed analysis is presented regarding the impact of decohering channels on thermal LQU. This controllable LQU in engineering applications can disclose the advantage enabled in the superconducting charge qubits for designing quantum computers and quantum batteries.

Abstract: Quantum state preparation is a fundamental building block for various problems on a quantum computer. A non-unitary operator for that is designed to decay unwanted states contained in an initial state by introducing ancilla qubits, and it acts probabilistically on the initial state. In this study, we clarified that this probabilistic nature is a drag for quantum advantages: the probabilistic algorithms do not accelerate the computational speed over the classical ones. Combining quantum amplitude amplification (QAA) with multi-step probabilistic algorithms is proposed to address this drawback, leading to quadratic speedup and quantum advantages. We have also found that by the multi-step probabilistic method with QAA shows advantages than quantum phase estimation at the viewpoint of infidelity. We also demonstrated it to confirm the quadratic speedup, using a probabilistic imaginary-time evolution (PITE) method as an example.

Abstract: We study the properties of a driven cavity coupled to several qubits in the framework of a dissipative Jaynes-Cummings model. We show that the rotating wave approximation (RWA) allows to reduce the description of original driven model to an effective Jaynes-Cummings model with strong coupling between photons and qubits. Two semi-analytical approaches are developed to describe the steady state of this system. We first treat the weak dissipation limit where we derive perturbative series of rate equations that converge to the exact RWA steady-state except near the cavity resonance. This approach exactly describes the multi-photon resonances in the system. Then in the strong dissipation limit we introduce a semiclassical approximation which allows to reproduce the mean spin-projections and cavity state. This approach reproduces the RWA exactly in the strong dissipation limit but provides good qualitative trends even in more quantum regimes. We then focus on quantum synchronization of qubits through their coupling to the cavity. We demonstrate the entangled steady state of a pair of qubits synchronized through their interaction with a driven cavity in presence of dissipation and decoherence. Finally we discuss synchronization of a larger number of qubits.

Abstract: We propose a novel variational ansatz for the ground state preparation of the $\mathbb{Z}_2$ lattice gauge theory (LGT) in quantum simulators (QSs). It combines dissipative and unitary operations in a completely deterministic scheme with a circuit complexity that does not scale with the size of the considered lattice. We find that, with very few variational parameters, the ansatz is able to achieve $>\!99\%$ fidelity with the true ground state in both the confined and deconfined phase of the $\mathbb{Z}_2$ LGT. We benchmark our proposal against the unitary Hamiltonian variational ansatz (HVA), and find a clear advantage of our scheme, especially for few variational parameters as well as for large system sizes. After performing a finite-size scaling analysis, we show that our dissipative variational ansatz is able to predict critical exponents with accuracies that surpass the capabilities of the HVA. Furthermore, we investigate the ground-state preparation algorithm in the presence of circuit-level noise and determine variational error thresholds, which determine error rates $p_{L}$, below which it would be beneficial to increase the number of layers $L \mapsto L+1$. Comparing those values to quantum gate errors $p$ of state-of-the-art quantum processors, we provide a detailed assessment of the prospects of our scheme to explore the $\mathbb{Z}_2$ LGT on near-term devices.

Abstract: A quantum engine with n qubits performing thermodynamic cycles with two thermal reservoirs is presented. While such constructions have been aplenty, here we show the existence of what we term as "limit cycle" at a purely quantum level of description owing to the properties of superoperators governing the evolution of states. It is shown that the limit cycle is the same under forward and reverse protocol of cycle operations. This limit cycle becomes the basis of the quantum engine. One dimensional Ising model has been used to illustrate these ideas.

Abstract: We introduce an algorithm to compute Hamiltonian dynamics on digital quantum computers that requires only a finite circuit depth to reach an arbitrary precision, i.e. achieves zero Trotter error with finite depth. This finite number of gates comes at the cost of an attenuation of the measured expectation value by a known amplitude, requiring more shots per circuit. The algorithm generalizes to time-dependent Hamiltonians, for example for adiabatic state preparation. This makes it particularly suitable for present-day relatively noisy hardware that supports only circuits with moderate depth.

Abstract: Quantum many-body interactions can induce quantum entanglement among particles, rendering them valuable resources for quantum-enhanced sensing. In this work, we derive a universal and fundamental bound for the growth of the quantum Fisher information. We apply our bound to the metrological protocol requiring only separable initial states, which can be readily prepared in experiments. By establishing a link between our bound and the Lieb-Robinson bound, which characterizes the operator growth in locally interacting quantum many-body systems, we prove that the precision cannot surpass the shot noise limit at all times in locally interacting quantum systems. This conclusion also holds for an initial state that is the non-degenerate ground state of a local and gapped Hamiltonian. These findings strongly hint that when one can only prepare separable initial states, nonlocal and long-range interactions are essential resources for surpassing the shot noise limit. This observation is confirmed through numerical analysis on the long-range Ising model. Our results bridge the field of many-body quantum sensing and operator growth in many-body quantum systems and open the possibility to investigate the interplay between quantum sensing and control, many-body physics and information scrambling

Abstract: With the rise of quantum technologies, data security increasingly relies on quantum cryptography and its most notable application, quantum key distribution (QKD). Yet, current technological limitations, in particular, the unavailability of quantum repeaters, cause relatively low key distribution rates in practical QKD implementations. Here, we demonstrate a remarkable improvement in the QKD performance using end-to-end line tomography for the wide class of relevant protocols. Our approach is based on the real-time detection of interventions in the transmission channel, enabling an adaptive response that modifies the QKD setup and post-processing parameters, leading, thereby, to a substantial increase in the key distribution rates. Our findings provide everlastingly secure efficient quantum cryptography deployment potentially overcoming the repeaterless rate-distance limit.

Abstract: We construct families of Floquet codes derived from colour code tilings of closed hyperbolic surfaces. These codes have weight-two check operators, a finite encoding rate and can be decoded efficiently with minimum-weight perfect matching. We also construct semi-hyperbolic Floquet codes, which have improved distance scaling, and are obtained via a fine-graining procedure. Using a circuit-based noise model that assumes direct two-qubit measurements, we show that semi-hyperbolic Floquet codes can be $48\times$ more efficient than planar honeycomb codes and therefore over $100\times$ more efficient than alternative compilations of the surface code to two-qubit measurements, even at physical error rates of $0.3\%$ to $1\%$. We further demonstrate that semi-hyperbolic Floquet codes can have a teraquop footprint of only 32 physical qubits per logical qubit at a noise strength of $0.1\%$. For standard circuit-level depolarising noise at $p=0.1\%$, we find a $30\times$ improvement over planar honeycomb codes and a $5.6\times$ improvement over surface codes. Finally, we analyse small instances that are amenable to near-term experiments, including a 16-qubit Floquet code derived from the Bolza surface.

Abstract: We investigate many-body dynamics where the evolution is governed by unitary circuits through the lens of `Krylov complexity', a recently proposed measure of complexity and quantum chaos. We extend the formalism of Krylov complexity to unitary circuit dynamics and focus on Floquet circuits arising as the Trotter decomposition of Hamiltonian dynamics. For short Trotter steps the results from Hamiltonian dynamics are recovered, whereas a large Trotter step results in different universal behavior characterized by the existence of local maximally ergodic operators: operators with vanishing autocorrelation functions, as exemplified in dual-unitary circuits. These operators exhibit maximal complexity growth, act as a memoryless bath for the dynamics, and can be directly probed in current quantum computing setups. These two regimes are separated by a crossover in chaotic systems. Conversely, we find that free integrable systems exhibit a nonanalytic transition between these different regimes, where maximally ergodic operators appear at a critical Trotter step.

Abstract: Building efficient large-scale quantum computers is a significant challenge due to limited qubit connectivities and noisy hardware operations. Transpilation is critical to ensure that quantum gates are on physically linked qubits, while minimizing $\texttt{SWAP}$ gates and simultaneously finding efficient decomposition into native $\textit{basis gates}$. The goal of this multifaceted optimization step is typically to minimize circuit depth and to achieve the best possible execution fidelity. In this work, we propose $\textit{MIRAGE}$, a collaborative design and transpilation approach to minimize $\texttt{SWAP}$ gates while improving decomposition using $\textit{mirror gates}$. Mirror gates utilize the same underlying physical interactions, but when their outputs are reversed, they realize a different or $\textit{mirrored}$ quantum operation. Given the recent attention to $\sqrt{\texttt{iSWAP}}$ as a powerful basis gate with decomposition advantages over $\texttt{CNOT}$, we show how systems that implement the $\texttt{iSWAP}$ family of gates can benefit from mirror gates. Further, $\textit{MIRAGE}$ uses mirror gates to reduce routing pressure and reduce true circuit depth instead of just minimizing $\texttt{SWAP}$s. We explore the benefits of decomposition for $\sqrt{\texttt{iSWAP}}$ and $\sqrt[4]{\texttt{iSWAP}}$ using mirror gates, including both expanding Haar coverage and conducting a detailed fault rate analysis trading off circuit depth against approximate gate decomposition. We also describe a novel greedy approach accepting mirror substitution at different aggression levels within MIRAGE. Finally, for $\texttt{iSWAP}$ systems that use square-lattice topologies, $\textit{MIRAGE}$ provides an average of 29.6\% reduction in circuit depth by eliminating an average of 59.9\% $\texttt{SWAP}$ gates, which ultimately improves the practical applicability of our algorithm.

Abstract: We introduce the problem of verification of stable sources in quantum systems. This problem is closely related to the problem of quantum verification first proposed by Pallister et. al. [1], however it extends the notion of the original problem. We introduce a family of states that come from a non-i.i.d. source which we call a Markov state. We prove in theorem 1 that these states are not well described with tensor products over a changing source. In theorem 2 we further provide a lower bound on the trace distance between two Markov states, which is the simplest way to solve the problem of verification of quantum stable sources.

Abstract: Understanding the dynamics of bound state formation is one of the fundamental questions in confining quantum field theories such as Quantum Chromodynamics (QCD). One hadronization mechanism that has garnered significant attention is the breaking of a string initially connecting a fermion and an anti-fermion. Deepening our understanding of real-time string-breaking dynamics with simpler, lower dimensional models like the Schwinger model can improve our understanding of the hadronization process in QCD and other confining systems found in condensed matter and statistical systems. In this paper, we consider the string-breaking dynamics within the Schwinger model and investigate its modification inside a thermal medium, treating the Schwinger model as an open quantum system coupled to a thermal environment. Within the regime of weak coupling between the system and environment, the real-time evolution of the system can be described by a Lindblad evolution equation. We analyze the Liouvillian gaps of this Lindblad equation and the time dependence of the system's von Neumann entropy. We observe that the late-time relaxation rate decreases as the environment correlation length increases. Moreover, when the environment correlation length is infinite, the system exhibits two steady states, one in each of the sectors with definite charge-conjugation-parity (CP) quantum numbers. For parameter regimes where an initial string breaks in vacuum, we observe a delay of the string breaking in the medium, due to kinetic dissipation effects. Conversely, in regimes where an initial string remains intact in vacuum time evolution, we observe string breaking (melting) in the thermal medium. We further discuss how the Liouvillian dynamics of the open Schwinger model can be simulated on quantum computers and provide an estimate of the associated Trotter errors.

Abstract: When estimating the eigenvalues of a given observable, even fault-tolerant quantum computers will be subject to errors, namely algorithmic errors. These stem from approximations in the algorithms implementing the unitary passed to phase estimation to extract the eigenvalues, e.g. Trotterisation or qubitisation. These errors can be tamed by increasing the circuit complexity, which may be unfeasible in early-stage fault-tolerant devices. Rather, we propose in this work an error mitigation strategy that enables a reduction of the algorithmic errors up to any order, at the cost of evaluating the eigenvalues of a set of observables implementable with limited resources. The number of required observables is estimated and is shown to only grow polynomially with the number of terms in the Hamiltonian, and in some cases, linearly with the desired order of error mitigation. Our results show error reduction of several orders of magnitude in physically relevant cases, thus promise accurate eigenvalue estimation even in early fault-tolerant devices with limited number of qubits.

Abstract: We propose and analyze a protocol for stabilizing a maximally entangled state of two noninteracting qubits using active state-dependent feedback from a continuous two-qubit half-parity measurement in coordination with a concurrent, non-commuting dynamical decoupling drive. We demonstrate the surprising result that such a drive be simultaneous with the measurement and feedback, and can also be part of the feedback protocol itself. We show that robust stabilization with near-unit fidelity can be achieved even in the presence of realistic nonidealities, such as time delay in the feedback loop, imperfect state-tracking, inefficient measurements, and dephasing from 1/f-distributed qubit-frequency noise. We mitigate feedback-delay error by introducing a forward-state-estimation strategy in the feedback controller that tracks the effects of control signals already in transit.

Abstract: Nearest-neighbour clustering is a simple yet powerful machine learning algorithm that finds natural application in the decoding of signals in classical optical fibre communication systems. Quantum nearest-neighbour clustering promises a speed-up over the classical algorithms, but the current embedding of classical data introduces inaccuracies, insurmountable slowdowns, or undesired effects. This work proposes the generalised inverse stereographic projection into the Bloch sphere as an encoding for quantum distance estimation in k nearest-neighbour clustering, develops an analogous classical counterpart, and benchmarks its accuracy, runtime and convergence. Our proposed algorithm provides an improvement in both the accuracy and the convergence rate of the algorithm. We detail an experimental optic fibre setup as well, from which we collect 64-Quadrature Amplitude Modulation data. This is the dataset upon which the algorithms are benchmarked. Through experiments, we demonstrate the numerous benefits and practicality of using the `quantum-inspired' stereographic k nearest-neighbour for clustering real-world optical-fibre data. This work also proves that one can achieve a greater advantage by optimising the radius of the inverse stereographic projection.

Abstract: We present a powerful formalism (d-CUT-E) to simulate the ultrafast quantum dynamics of molecular polaritons in the collective strong coupling regime, where a disordered ensemble of $N\gg10^{6}$ molecules couples to a cavity mode. Notably, we can capture this dynamics with a cavity hosting a single $\textit{effective}$ molecule with $\sim N_{bins}$ electronic states, where $N_{bins}\ll N$ is the number of bins discretizing the disorder distribution. Using d-CUT-E, we show that in highly disordered ensembles, total reaction yield upon broadband excitation converges to that outside of the cavity. Yet, strong coupling can bestow different reactivities upon individual molecules, leading to changes in reaction yield upon narrowband excitation. Crucially, this effect goes beyond changes in linear absorption due to optical filtering through polaritons.