arXiv subjects

Rawad Mezher

Publications and source records attributed to Rawad Mezher.

At least 19 recordsLinked to original sources

arXiv2026

Efficient training of photonic quantum generative models

The topic of generative learning has gained traction within the field of quantum machine learning, in particular with the advent of train-on-classical, deploy-on-quantum methods. This approach exploits the properties of intermediate-complexity circuits whose training can be simulated classically efficiently, but that generally require quantum hardware for the corresponding sampling problem. Quantum linear optics possess similar properties, which allow us to propose an efficient training procedure for photon-native quantum generative models based on the maximum mean discrepancy, where the deployment of the model corresponds to the task of boson sampling. We provide numerical results, propose datasets, and we also explore how initialization strategies, kernel and ansatz choice affect the training.

quant-ph↗

arXiv2026

Mitigating photon loss in linear optical quantum circuits

Photon loss rates set an effective upper limit on the size of computations that can be run on current linear optical quantum devices. We present a family of techniques designed to mitigate the effects of photon loss on both output probabilities and expectation values derived from noisy linear optical circuits composed of an input of n photons, an m-mode interferometer, and m single photon detectors. Central to these techniques is the construction recycled probabilities. Recycled probabilities are constructed from output statistics affected by loss, and are designed to amplify the signal of the ideal (lossless) probabilities. Classical postprocessing techniques then take recycled probabilities as input and output a set of loss-mitigated probabilities, or expectation values. Our postprocessing methods result in biased estimators of the lossless probabilities. Nevertheless, we provide both analytical and numerical evidence that these methods can be applied, up to large sample sizes, to produce output probabilities with lower combined bias and statistical errors than the statistical errors of the output probabilities obtained from postselection. Therefore, these methods can outperform postselection - currently the standard method of coping with photon loss when sampling from discrete variable linear optical quantum circuits. In contrast, we provide evidence that the popular zero-noise extrapolation technique cannot improve on the performance of postselection for any photon loss rate.

quant-ph↗

arXiv2026

Exponential improvement in benchmarking multiphoton interference

Several photonic quantum technologies rely on the ability to generate multiple indistinguishable photons. Benchmarking the level of indistinguishability of these photons is essential for scalability. The Hong-Ou-Mandel dip provides a benchmark for the indistinguishability between two photons, and extending this test to the multi-photon setting has so far resulted in a protocol that computes the genuine n-photon indistinguishability (GI). However, this protocol has a sample complexity that increases exponentially with the number of input photons for an estimation of GI up to a given additive error. To address this problem, we introduce new theorems that strengthen our understanding of the relationship between distinguishability and the suppression laws of the quantum Fourier transform interferometer (QFT). Building on this, we propose a protocol using the QFT for benchmarking GI that achieves constant sample complexity for the estimation of GI up to a given additive error for prime photon numbers, and sub-polynomial scaling otherwise, representing an exponential improvement over the state of the art. We prove the optimality of our protocol in many relevant scenarios and validate our approach experimentally on Quandela's reconfigurable photonic quantum processor, where we observe a clear advantage in runtime and precision over the state of the art. We therefore establish the first scalable method for computing multi-photon indistinguishability, which applies naturally to current and near-term photonic quantum hardware.

quant-ph↗

arXiv2025

Establishing Baselines for Photonic Quantum Machine Learning: Insights from an Open, Collaborative Initiative

The Perceval Challenge is an open, reproducible benchmark designed to assess the potential of photonic quantum computing for machine learning. Focusing on a reduced and hardware-feasible version of the MNIST digit classification task or near-term photonic processors, it offers a concrete framework to evaluate how photonic quantum circuits learn and generalize from limited data. Conducted over more than three months, the challenge attracted 64 teams worldwide in its first phase. After an initial selection, 11 finalist teams were granted access to GPU resources for large-scale simulation and photonic hardware execution through cloud service. The results establish the first unified baseline of photonic machine-learning performance, revealing complementary strengths between variational, hardware-native, and hybrid approaches. This challenge also underscores the importance of open, reproducible experimentation and interdisciplinary collaboration, highlighting how shared benchmarks can accelerate progress in quantum-enhanced learning. All implementations are publicly available in a single shared repository (https://github.com/Quandela/HybridAIQuantum-Challenge), supporting transparent benchmarking and cumulative research. Beyond this specific task, the Perceval Challenge illustrates how systematic, collaborative experimentation can map the current landscape of photonic quantum machine learning and pave the way toward hybrid, quantum-augmented AI workflows.

quant-ph↗

arXiv2025

On the role of coherence for quantum computational advantage

Quantifying the resources available to a quantum computer appears to be necessary to separate quantum from classical computation. Among them, entanglement, nonstabilizerness and coherence are arguably of great significance. We introduce path coherence as a measure of the coherent paths interferences arising in a quantum computation. Leveraging the sum-over-paths formalism, we obtain a classical algorithm for estimating quantum transition amplitudes, the complexity of which scales with path coherence. As path coherence relates to the hardness of classical estimation of quantum transition amplitudes, it provides a new perspective on the role of coherence in quantum computational advantage. Beyond their fundamental significance, our results have practical applications for simulating large classes of quantum computations with classical computers.

quant-ph↗

arXiv2025

New designs of linear optical interferometers with minimal depth and component count

We adapt an algorithm for CNOT circuits synthesis based on the Bruhat decomposition to the design of linear optical circuits with Mach-Zehnder interferometers (MZI). The synthesis algorithm reduces to designing sorting networks with nearest neighbor swapping operations as elementary gates. We recover previous designs from the literature but with additional theoretical properties regarding the compiler that implements unitaries on the interferometer. Notably the compiler can always decide whether a unitary can be implemented on a given interferometer and, if so, returns the shallowest possible implementation. We also show natural extensions of our framework for boson sampling experiments and for the coupling of multiple integrated interferometers to design larger linear optical systems. In both cases, the designs are optimal in terms of number of optical components. Finally, we propose a greedy design which exploits the arbritrary-but-fixed coupling of separate integrated interferometers to perform shallow boson sampling. We discuss the optimal interferometer dimensions to maximize the transmission. Beyond boson sampling, our developed framework allows a resource-favourable implemention of any non-adaptive linear optical quantum algorithm, by providing the shallowest possible interferometer for implementing this algorithm.

quant-ph↗

arXiv2025

Error-mitigated photonic quantum circuit Born machine

In this article, we study quantum circuit Born machines (QCBMs) in the context of photonic quantum computing. QCBMs are a popular choice of quantum generative machine learning models, and we present a QCBM designed for linear optics. We show that a recently developed error mitigation technique called recycling mitigation greatly improves the training of QCBMs in realistic scenarios with photon loss, which is the primary source of noise in photonic systems. We demonstrate this through numerical simulations and through an experiment on a quantum photonic integrated processor. We expect our work to pave the way towards more demonstrations of error mitigation techniques tailored to photonic devices which can enhance the performance of a quantum algorithm.

quant-ph↗

arXiv2024

Connecting quantum circuit amplitudes and matrix permanents through polynomials

In this paper, we strengthen the connection between qubit-based quantum circuits and photonic quantum computation. Within the framework of circuit-based quantum computation, the sum-over-paths interpretation of quantum probability amplitudes leads to the emergence of sums of exponentiated polynomials. In contrast, the matrix permanent is a combinatorial object that plays a crucial role in photonic by describing the probability amplitudes of linear optical computations. To connect the two, we introduce a general method to encode an $\mathbb F_2$-valued polynomial with complex coefficients into a graph, such that the permanent of the resulting graph's adjacency matrix corresponds directly to the amplitude associated the polynomial in the sum-over-path framework. This connection allows one to express quantum amplitudes arising from qubit-based circuits as permanents, which can naturally be estimated on a photonic quantum device.

quant-ph↗

arXiv2024

Restricted Randomized Benchmarking with Universal Gates of Fixed Sequence Length

The standard randomized benchmarking protocol requires access to often complex operations that are not always directly accessible. Compiler optimization does not always ensure equal sequence length of the directly accessible universal gates for each random operation. We introduce a version of the RB protocol that creates Haar-randomness using a directly accessible universal gate set of equal sequence length rather than relying upon a t-design or even an approximate one. This makes our protocol highly resource efficient and practical for small qubit numbers. We exemplify our protocol for creating Haar-randomness in the case of single and two qubits. Benchmarking our result with the standard RB protocol, allows us to calculate the overestimation of the average gate fidelity as compared to the standard technique. We augment our findings with a noise analysis which demonstrates that our method could be an effective tool for building accurate models of experimental noise.

quant-ph↗

arXiv2023

Solving graph problems with single-photons and linear optics

An important challenge for current and near-term quantum devices is finding useful tasks that can be preformed on them. We first show how to efficiently encode a bounded $n \times n$ matrix $A$ into a linear optical circuit with $2n$ modes. We then apply this encoding to the case where $A$ is a matrix containing information about a graph $G$. We show that a photonic quantum processor consisting of single-photon sources, a linear optical circuit encoding $A$, and single-photon detectors can solve a range of graph problems including finding the number of perfect matchings of bipartite graphs, computing permanental polynomials, determining whether two graphs are isomorphic, and the $k$-densest subgraph problem. We also propose pre-processing methods to boost the probabilities of observing the relevant detection events and thus improve performance. Finally we present both numerical simulations and implementations on Quandela's Ascella photonic quantum processor to validate our findings.

quant-ph↗

arXiv2023

A general-purpose single-photon-based quantum computing platform

Quantum computing aims at exploiting quantum phenomena to efficiently perform computations that are unfeasible even for the most powerful classical supercomputers. Among the promising technological approaches, photonic quantum computing offers the advantages of low decoherence, information processing with modest cryogenic requirements, and native integration with classical and quantum networks. To date, quantum computing demonstrations with light have implemented specific tasks with specialized hardware, notably Gaussian Boson Sampling which permitted quantum computational advantage to be reached. Here we report a first user-ready general-purpose quantum computing prototype based on single photons. The device comprises a high-efficiency quantum-dot single-photon source feeding a universal linear optical network on a reconfigurable chip for which hardware errors are compensated by a machine-learned transpilation process. Our full software stack allows remote control of the device to perform computations via logic gates or direct photonic operations. For gate-based computation we benchmark one-, two- and three-qubit gates with state-of-the art fidelities of $99.6\pm0.1 \%$, $93.8\pm0.6 \%$ and $86\pm1.2 \%$ respectively. We also implement a variational quantum eigensolver, which we use to calculate the energy levels of the hydrogen molecule with high accuracy. For photon native computation, we implement a classifier algorithm using a $3$-photon-based quantum neural network and report a first $6$-photon Boson Sampling demonstration on a universal reconfigurable integrated circuit. Finally, we report on a first heralded 3-photon entanglement generation, a key milestone toward measurement-based quantum computing.

quant-ph↗

arXiv2023

Perceval: A Software Platform for Discrete Variable Photonic Quantum Computing

We introduce Perceval, an open-source software platform for simulating and interfacing with discrete-variable photonic quantum computers, and describe its main features and components. Its Python front-end allows photonic circuits to be composed from basic photonic building blocks like photon sources, beam splitters, phase-shifters and detectors. A variety of computational back-ends are available and optimised for different use-cases. These use state-of-the-art simulation techniques covering both weak simulation, or sampling, and strong simulation. We give examples of Perceval in action by reproducing a variety of photonic experiments and simulating photonic implementations of a range of quantum algorithms, from Grover's and Shor's to examples of quantum machine learning. Perceval is intended to be a useful toolkit for experimentalists wishing to easily model, design, simulate, or optimise a discrete-variable photonic experiment, for theoreticians wishing to design algorithms and applications for discrete-variable photonic quantum computing platforms, and for application designers wishing to evaluate algorithms on available state-of-the-art photonic quantum computers.

quant-ph↗

arXiv2022

Mitigating errors by quantum verification and post-selection

Correcting errors due to noise in quantum circuits run on current and near-term quantum hardware is essential for any convincing demonstration of quantum advantage. Indeed, in many cases it has been shown that noise renders quantum circuits efficiently classically simulable, thereby destroying any quantum advantage potentially offered by an ideal (noiseless) implementation of these circuits. Although the technique of quantum error correction (QEC) allows to correct these errors very accurately, QEC usually requires a large overhead of physical qubits which is not reachable with currently available quantum hardware. This has been the motivation behind the field of quantum error mitigation, which aims at developing techniques to correct an important part of the errors in quantum circuits, while also being compatible with current and near-term quantum hardware. In this work, we present a technique for quantum error mitigation which is based on a technique from quantum verification, the so-called accreditation protocol, together with post-selection. Our technique allows for correcting the expectation value of an observable $O$, which is the output of multiple runs of noisy quantum circuits, where the noise in these circuits is at the level of preparations, gates, and measurements. We discuss the sample complexity of our procedure and provide rigorous guarantees of errors being mitigated under some realistic assumptions on the noise. Our technique also allows for time dependant behaviours, as we allow for the output states to be different between different runs of the accreditation protocol. We validate our findings by running our technique on currently available quantum hardware.

quant-ph↗

arXiv2022

Assessing the quality of near-term photonic quantum devices

For near-term quantum devices, an important challenge is to develop efficient methods to certify that noise levels are low enough to allow potentially useful applications to be carried out. We present such a method tailored to photonic quantum devices consisting of single photon sources coupled to linear optical circuits coupled to photon detectors. It uses the output statistics of BosonSampling experiments with input size $n$ ($n$ input photons in the ideal case). We propose a series of benchmark tests targetting two main sources of noise, namely photon loss and distinguishability. Our method results in a single-number metric, the Photonic Quality Factor, defined as the largest number of input photons for which the output statistics pass all tests. We provide strong evidence that passing all tests implies that our experiments are not efficiently classically simulable, by showing how several existing classical algorithms for efficiently simulating noisy BosonSampling fail the tests. Finally we show that BosonSampling experiments with average photon loss rate per mode scaling as $o(1)$ and average fidelity of $ (1-o(\frac{1}{n^6}))^2$ between any two single photon states is sufficient to keep passing our tests. Unsurprisingly, our results highlight that scaling in a manner that avoids efficient classical simulability will at some point necessarily require error correction and mitigation.

quant-ph↗

arXiv2021

Randomized Benchmarking with Stabilizer Verification and Gate Synthesis

Recently, there has been an emergence of useful applications for noisy intermediate-scale quantum (NISQ) devices notably, though not exclusively, in the fields of quantum machine learning and variational quantum algorithms. In such applications, circuits of various depths and composed of different sets of gates are run on NISQ devices. Therefore, it is crucial to find practical ways to capture the general performance of circuits on these devices. Motivated by this pressing need, we modified the standard Clifford randomized benchmarking (RB) and interleaved RB schemes targeting them to hardware limitations. Firstly we remove the requirement for, and assumptions on, the inverse operator, in Clifford RB by incorporating a tehchnique from quantum verification. This introduces another figure of merit by which to assess the quality of the NISQ hardware, namely the acceptance probability of quantum verification. Many quantum algorithms, that provide an advantage over classical algorithms, demand the use of Clifford as well as non-Clifford gates. Therefore, as our second contribution we develop a technique for characterising a variety of non-Clifford gates, by combining tools from gate synthesis with interleaved RB. Both of our techniques are most relevant when used in conjunction with RB schemes that benchmark generators (or native gates) of the Clifford group, and in low error regimes.

quant-ph↗

arXiv2021

Efficient Construction of Quantum Physical Unclonable Functions with Unitary t-designs

Quantum physical unclonable functions, or QPUFs, are rapidly emerging as theoretical hardware solutions to provide secure cryptographic functionalities such as key-exchange, message authentication, entity identification among others. Recent works have shown that in order to provide provable security of these solutions against any quantum polynomial time adversary, QPUFs are required to be a unitary sampled uniformly randomly from the Haar measure. This however is known to require an exponential amount of resources. In this work, we propose an efficient construction of these devices using unitary t-designs, called QPUF_t. Along the way, we modify the existing security definitions of QPUFs to include efficient constructions and showcase that QPUF_t still retains the provable security guarantees against a bounded quantum polynomial adversary with t-query access to the device. This also provides the first use case of unitary t-design construction for arbitrary t, as opposed to previous applications of t-designs where usually a few (relatively low) values of t are known to be useful for performing some task. We study the noise-resilience of QPUF_t against specific types of noise, unitary noise, and show that some resilience can be achieved particularly when the error rates affecting individual qubits become smaller as the system size increases. To make the noise-resilience more realistic and meaningful, we conclude that some notion of error mitigation or correction should be introduced.

quant-ph↗

arXiv2020

Fault-tolerant quantum speedup from constant depth quantum circuits

A defining feature in the field of quantum computing is the potential of a quantum device to outperform its classical counterpart for a specific computational task. By now, several proposals exist showing that certain sampling problems can be done efficiently quantumly, but are not possible efficiently classically, assuming strongly held conjectures in complexity theory. A feature dubbed quantum speedup. However, the effect of noise on these proposals is not well understood in general, and in certain cases it is known that simple noise can destroy the quantum speedup. Here we develop a fault-tolerant version of one family of these sampling problems, which we show can be implemented using quantum circuits of constant depth. We present two constructions, each taking $poly(n)$ physical qubits, some of which are prepared in noisy magic states. The first of our constructions is a constant depth quantum circuit composed of single and two-qubit nearest neighbour Clifford gates in four dimensions. This circuit has one layer of interaction with a classical computer before final measurements. Our second construction is a constant depth quantum circuit with single and two-qubit nearest neighbour Clifford gates in three dimensions, but with two layers of interaction with a classical computer before the final measurements. For each of these constructions, we show that there is no classical algorithm which can sample according to its output distribution in $poly(n)$ time, assuming two standard complexity theoretic conjectures hold. The noise model we assume is the so-called local stochastic quantum noise. Along the way, we introduce various new concepts such as constant depth magic state distillation (MSD), and constant depth output routing, which arise naturally in measurement based quantum computation (MBQC), but have no constant-depth analogue in the circuit model.

quant-ph↗

arXiv2020

Efficient approximate unitary t-designs from partially invertible universal sets and their application to quantum speedup

At its core a $t$-design is a method for sampling from a set of unitaries in a way which mimics sampling randomly from the Haar measure on the unitary group, with applications across quantum information processing and physics. We construct new families of quantum circuits on $n$-qubits giving rise to $\varepsilon$-approximate unitary $t$-designs efficiently in $O(n^3t^{12})$ depth. These quantum circuits are based on a relaxation of technical requirements in previous constructions. In particular, the construction of circuits which give efficient approximate $t$-designs by Brandao, Harrow, and Horodecki (F.G.S.L Brandao, A.W Harrow, and M. Horodecki, Commun. Math. Phys. (2016).) required choosing gates from ensembles which contained inverses for all elements, and that the entries of the unitaries are algebraic. We reduce these requirements, to sets that contain elements without inverses in the set, and non-algebraic entries, which we dub partially invertible universal sets. We then adapt this circuit construction to the framework of measurement based quantum computation(MBQC) and give new explicit examples of $n$-qubit graph states with fixed assignments of measurements (graph gadgets) giving rise to unitary $t$-designs based on partially invertible universal sets, in a natural way. We further show that these graph gadgets demonstrate a quantum speedup, up to standard complexity theoretic conjectures. We provide numerical and analytical evidence that almost any assignment of fixed measurement angles on an $n$-qubit cluster state give efficient $t$-designs and demonstrate a quantum speedup.

quant-ph↗