Abstract
Quantum state tomography (QST) via local measurements on reduced density matrices (LQST) is a promising approach but becomes impractical for large systems. To tackle this challenge, we developed an efficient quantum state tomography method inspired by quantum overlapping tomography, which utilizes parallel measurements (PQST). In contrast to LQST, PQST significantly reduces the number of measurements and offers more robustness against shot noise. Experimentally, we demonstrate the feasibility of PQST in a tree-like superconducting qubit chip by designing high-efficiency circuits, preparing W states, ground states of Hamiltonians and random states, and then reconstructing these density matrices using full quantum state tomography (FQST), LQST, and PQST. Our results show that PQST reduces measurement cost, achieving fidelities of 98.68% and 95.07% after measuring 75 and 99 observables for 6-qubit and 9-qubit W states, respectively. Furthermore, the reconstruction of the largest density matrix of the 12-qubit W state is achieved with the similarity of 89.23% after just measuring 243 parallel observables, while $3^{12}=531441$ complete observables are needed for FQST.
Introduction
As quantum processors scale to dozens or hundreds of qubits, a major "characterization bottleneck" has emerged. Verifying the state of a quantum system traditionally requires Full Quantum State Tomography (FQST), which scales exponentially with the number of qubits ($N$). For $N$ qubits, $3^N$ different observables must be measured. While a 10-qubit system might take five days of continuous data collection, a 12-qubit system ($3^{12} = 531,441$ observables) becomes practically impossible, requiring over 60 days of measurement time.
To address this, we present Parallel Quantum State Tomography (PQST). This method leverages overlapping measurements and tensor network representations to reconstruct quantum states with massive speedups, bridging the gap between building large-scale quantum hardware and the ability to verify the resulting states in a reasonable timeframe.
Methods and Experimental Implementation
The PQST framework integrates three core innovations to achieve sample efficiency and noise robustness:
- Parallel Overlapping Measurements: Utilizing "Perfect Hash Functions" (FPHF), qubits are grouped to allow simultaneous measurement in different bases. This ensures that all $k$-qubit correlations (Reduced Density Matrices or RDMs) are captured within a minimal set of parallel measurement settings.
- Locally Purified States (LPS): To manage the memory requirements of large density matrices, the state is represented as a tensor network (LPS). This approach is particularly effective for the "mixed" states typical of noisy experimental hardware.
- Maximum Likelihood Estimation (MLE): Instead of relying on average expectation values, raw "shot" data is processed through an MLE loss function, enhancing robustness against statistical fluctuations (shot noise).
The experiments were conducted on a 31-transmon superconducting qubit chip with a tree-like topology. The team prepared several benchmark states, including highly entangled W states ($|W_N\rangle$ for $N=6, 9, 12$), ground states of fully-connected Hamiltonians, and random states generated by quantum circuits.
Results and Discussion
The performance of PQST was benchmarked against standard FQST and Local QST (LQST). The results demonstrate a significant reduction in measurement overhead without sacrificing accuracy. For a 12-qubit W state, PQST achieved a reconstruction with 89.23% similarity using only 243 parallel observables. This represents a reduction of several orders of magnitude compared to the 531,441 observables required for FQST, reducing the time from months to minutes.
A key insight from the study is the "information recycling" inherent in parallel measurements. Data collected for one set of qubits automatically informs the characterization of other overlapping subsystems. Furthermore, PQST proved more robust to shot noise than LQST; at equivalent total sample sizes, PQST consistently yielded higher fidelity. The method also accurately predicted unmeasured subsystem properties, such as entanglement negativity, matching the results of much more expensive FQST.
Conclusion
Parallel Quantum State Tomography offers a scalable path for characterizing large-scale quantum devices. While the current implementation focused on tree-like topologies and specific state classes, the underlying principles of parallelized measurement and tensor network reconstruction are broadly applicable. PQST stands as a vital tool for the future benchmarking, characterization, and property learning of complex quantum systems.
Summary updated