Western Australia Team Hits 92·9% Fidelity in State Prep

Quantum state preparation with rapid, strong fidelity verification is now possible on near-term quantum hardware. This milestone was realised without relying on idealised assumptions or deep fault-tolerant overhead, but through resource-minimal circuits optimised for NISQ-era and early fault-tolerant devices. Furthermore, a key limitation in current quantum state certification has been addressed. Validation methods like shadow overlap work well for random states, yet their sample complexity scales poorly with structured states.

Demonstrating efficient quantum validation via reduced measurement requirements and digitised signal encoding

A reduction exceeding ten orders of magnitude in the verification parameter τ was achieved, falling from levels previously hindering certification to below 0.929; this enables robust validation using only 1,000 quantum measurement shots. Scientists at The University of Western Australia assessed end-to-end fidelity on Quantinuum’s H2-1 trapped-ion platform by combining resource-efficient tensor network based state preparation with a novel change-of-basis technique within shadow overlap methodology, without relying upon fault tolerance or idealised assumptions. Conventional techniques demanded exponentially more measurements rendering them impractical for near-term devices and thus unable to reliably verify highly structured states essential for advanced algorithms.

Thirteen qubits were used to successfully encode and validate a digitised acoustic signal onto the H2-1 trapped-ion platform; this required a quantum circuit comprising 229 single-qubit gates alongside 108 two-qubit gates demonstrating practical scalability. Optimised circuits suited to current technology allowed a hardware fidelity of 0.929 to be accomplished without employing fault tolerance techniques. The team also demonstrated that their new change-of-basis technique reduces the computational burden of verifying these structured states by over ten orders of magnitude, enabling strong certification with 1,000 measurement shots.

Reducing Verification Overhead via Pre-measurement Basis Transformations

Addressing high computational complexity when dealing with structured quantum states used in practical algorithms is increasingly important for scientists; they achieve this through introducing a pre-measurement basis-change technique which lowers τ for such targets. This approach tightens theoretical certification guarantees of shadow overlap and integrates tensor network preparation and validation into a unified workflow. Consequently, how classical data can be mapped to and verified on quantum hardware under realistic noise and measurement budgets has shifted.

The framework offers an immediate, scalable benchmarking standard by compressing verification to only 1,000 measurements. Accurate state preparation is fundamental for implementing quantum image processing, machine learning, linear algebra, and chemistry algorithms; maximising accuracy while minimising total resources remains key. Validating fidelity after preparing states on imperfect computers presents challenges because accurate preparation is imperative for downstream tasks representing bottlenecks in realising experimental quantum algorithms.

Scientists propose an end-to-end scheme aimed at minimising resource overhead which was realised using Quantinuum’s H2-1 trapped-ion computer. Recent advances position tensor networks as leading candidates for efficient and flexible state preparation with minimal overhead due to their limited requirements and shallow circuits suiting near-term implementation including existing hardware. Tensor networks efficiently represent structured classical data allowing target states to be computed via tensor cross interpolation even when dealing with large numbers of qubits.

For experimental realisation, they employ a methodology based on classically optimised disentangling preparing matrix product states through shallow quantum circuits consisting primarily of 2-qubit gates. This design explicitly minimises circuit depth without ancillas making it promising for the near-term preparation of both classical data and other target states. A primary challenge in experimentally realising quantum state preparation lies in validating successful encoding; an ideal certification procedure uses only local measurements but probing global entanglement structure proves difficult given limited qubit locality.

Previous methods often required deep circuits or were restricted to specific target states. Reliable preparation remains essential for quantum computation while full tomography becomes impractical meaning that validating prepared state fidelity is important. These challenges hindered many techniques however Huang et al. developed “shadow overlap” which aimed to address these issues. Although τ scales efficiently for most states, its effectiveness relies on this parameter remaining manageable, a value it can become prohibitively large when applied to structured states used in practical algorithms.

A pre-measurement basis change technique was introduced reducing τ by over ten orders of magnitude for such targets. A Haar random quantum state possesses different statistical properties compared with those having specific structures like ground states satisfying area laws of entanglement entropy and direct computation reveals high τ values rendering certification ineffective theoretically. Scientists overcame this limitation introducing a change-of-basis stage prior to using shadow overlap; this reduced τ by more than ten orders of magnitude as an optimal basis is computable for matrix product states making it compatible with tensor network preparation.

To demonstrate their approach empirically, they prepared a segment of acoustic signal on 13 qubits via circuits consisting of 229 single and 108 two qubit gates certifying success utilising shadow overlap while decreasing uncertainty in measured fidelity. This structure underpins both state preparation method efficiency and τ reduction layer computation. The Schmidt Spectrum Optimisation (SSO) algorithm constructs shallow circuits creating approximate MPS representations of n-qubit pure states.

Validating single acoustic signals unlocks nearer term quantum device assessment

While validating quantum states is now possible without relying on overly complex circuits benefiting near-term devices this demonstration currently focuses solely upon encoding a single digitised acoustic signal. This raises questions about scalability; can the technique be readily extended to handle more substantial datasets or diverse types of structured information vital for real-world applications like drug discovery and materials science. The paper acknowledges that assessing performance with different data remains an open challenge hinting at potential bottlenecks as complexity increases.

Despite extending verification to larger datasets presenting hurdles, it does not diminish its present value. The team has established a complete system for preparing and verifying quantum states on real-world hardware moving beyond demonstrating preparation to confirming accuracy under realistic conditions. Achieving high fidelity without relying upon complex error correction is particularly noteworthy given current limitations in building perfect computers. This streamlined approach compresses strong verification down to 1,000 measurements offering an immediate benchmarking standard applicable across various near-term devices.

Researchers successfully prepared and verified a structured quantum state encoding a digitised acoustic signal on the Quantinuum H2-1 trapped-ion platform with a hardware fidelity of 0.929. This demonstrates that accurate quantum state validation is possible using resource-efficient circuits suitable for today’s early quantum computing systems. The authors suggest further work will focus on assessing performance when applied to different datasets and scaling up this approach.

More like this

Quantum Error CorrectionBBVA, FirstQFM, Florida Atlantic University and Jülich Get Early Access to D-Wave’s Gate Simulator

Quantum Error CorrectionIonQ’s quantum qLDPC code boosts Toffoli gates 5.0x

Quantum HardwareNew design removes 98.4% of qubit leakage during measurement

Quantum Error CorrectionResearchers Achieve Stable Quantum Memory at Error Rates below Zero Point Four Per Cent