Quantum error correction is essential for advanced quantum algorithms, but it is debated whether current processors can run shorter-depth quantum circuits at a scale that provides practical advantages. Researchers used a superconducting quantum processor with 127 qubits to run quantum circuits with up to 60 layers of two-qubit gates, totalling 2,880 CNOT gates. These circuits are beyond what is feasible with brute-force classical methods. The study demonstrates the ability of existing devices to perform accurate computations at a scale beyond classical simulation, paving the way for further research into deriving practical computational advantages from noise-limited quantum circuits.
Introduction
IBM researchers have used a superconducting quantum processor with 127 qubits to run quantum circuits with up to 60 layers of two-qubit gates, totalling 2,880 CNOT gates. This achievement demonstrates the ability of existing devices to perform accurate computations at a scale beyond brute-force classical simulation. The results suggest that there is merit in pursuing research towards deriving a practical computational advantage from noise-limited quantum circuits, even before the advent of fault-tolerant quantum computing.
Quantum Error Correction
Quantum error correction is essential for advanced quantum algorithms, but it is debated whether current processors can run shorter-depth quantum circuits at a scale that could provide practical advantages. Despite the progress of quantum hardware, simple fidelity bounds support this pessimistic forecast. However, using classical post-processing, the error-mitigation approach to near-term quantum advantage on noisy devices addresses this question by producing accurate expectation values from several different runs of the noisy quantum circuit.
Superconducting Quantum Processor
A superconducting quantum processor with 127 qubits runs quantum circuits with up to 60 layers of two-qubit gates, totalling 2,880 CNOT gates. General quantum circuits of this size are beyond what is feasible with brute-force classical methods. The benchmark circuit is the Trotterized time evolution of a 2D transverse-field Ising model, which has found creative extensions in recent simulations exploring quantum many-body phenomena.
Noise Characterization and Scaling
Noise characterization and scaling for 127-qubit Trotterized time-evolution circuits are essential for understanding the performance of quantum processors. Probabilistic error cancellation (PEC) effectively provides unbiased estimates of observables. However, for the current error rates on the device, the sampling overhead for the circuit volumes considered remains restrictive. Zero-noise extrapolation (ZNE) provides a biased estimator at a potentially much lower sampling cost and has been widely adopted due to its simplicity and effectiveness.
Tensor Network Simulations
Tensor networks have been widely used to approximate and compress quantum state vectors that arise in the study of low-energy eigenstates and time evolution by local Hamiltonians. They have also been successfully used to simulate low-depth noisy quantum circuits. Simulation accuracy can be improved by increasing the bond dimension, which constrains the amount of entanglement of the represented quantum state at a computational cost scaling polynomially with the bond dimension.
Quantum Processor Performance
The performance of quantum processors has improved significantly in recent years, enabling larger problems to be successfully executed with error mitigation. However, there is still room for improvement in gate fidelities and speed of superconducting quantum systems. As classical approximation methods advance and quantum hardware continue to improve, the utility of noisy quantum computers is expected to increase.
“The observation that a noisy quantum processor, even before the advent of fault-tolerant quantum computing, produces reliable expectation values at a scale beyond 100 qubits and non-trivial circuit depth leads to the conclusion that there is indeed merit to pursuing research towards deriving a practical computational advantage from noise-limited quantum circuits.”
Summary
Researchers have used a superconducting quantum processor with 127 qubits to run quantum circuits with up to 60 layers of two-qubit gates, demonstrating the ability of existing devices to perform accurate computations at a scale beyond brute-force classical simulation. This progress brings us closer to deriving practical computational advantages from noise-limited quantum circuits and motivates further advancements in both quantum and classical approximation methods.
- Quantum error correction is essential for advanced quantum algorithms, but it is debated whether current processors can run shorter-depth quantum circuits at a scale that provides practical advantages.
- Researchers used a superconducting quantum processor with 127 qubits to run quantum circuits with up to 60 layers of two-qubit gates, a total of 2,880 CNOT gates.
- The benchmark circuit is the Trotterized time evolution of a 2D transverse-field Ising model, which has applications in various areas of physics.
- The IBM Eagle processor ibm_kyiv, composed of 127 fixed-frequency transmon qubits, was used for the experimental implementation.
- The study demonstrates that noisy quantum processors can produce reliable expectation values at a scale beyond 100 qubits and non-trivial circuit depth.
- This research supports the pursuit of practical computational advantages from noise-limited quantum circuits and motivates the development of classical approximation methods.
Read More from the Article Published in Nature.
https://www.nature.com/articles/s41586-023-06096-3#Sec2
