Certifying almost all quantum states with few single-qubit measurements

Overview

• The paper presents a new technique for efficiently certifying that a quantum state synthesized in a lab is close to a target state, using only a small number of single-qubit measurements.
• This is an important task in quantum information science, but existing rigorous protocols require either deep quantum circuits or exponentially many measurements.
• The new technique relates certification to the mixing time of a random walk, allowing certification of almost all n-qubit target states, including those with exponential circuit complexity, from only O(n^2) single-qubit measurements.

Plain English Explanation

In the world of quantum computing, researchers often need to verify that a quantum state they've created in the lab is close to a target state they had in mind. This is an important task, but the existing rigorous methods for doing this either require very complex quantum circuits or an exponential number of individual qubit measurements, which can be impractical.

The researchers in this paper have found a new way to address this problem. Their key insight is that they can relate the task of verifying the quantum state to the behavior of a random walk. Using this connection, they show that for almost all possible target quantum states - even ones that would be extremely difficult to create using quantum circuits - they can verify the state using only a small number of simple measurements on individual qubits, specifically O(n^2) measurements for an n-qubit system.

This new efficient learning of quantum states prepared with few non-local resources technique has several important applications. It can be used to more easily benchmark quantum systems, optimize quantum circuits to generate desired target states, and verify representations of quantum states like neural networks, tensor networks, and other models. And the verified representations can then be used to efficiently predict hard-to-measure quantum properties that would otherwise require an exponential number of measurements.

The researchers demonstrate these applications in numerical experiments with up to 120 qubits, and show advantages over existing methods like cross-entropy benchmarking (XEB).

Technical Explanation

The key technical insight of the paper is that the task of certifying an n-qubit quantum state as close to a target state can be related to the mixing time of a random walk. Specifically, the researchers show that almost all n-qubit target states, including those with exponential circuit complexity, can be certified from only O(n^2) single-qubit measurements.

This is established through a new protocol that uses the random walk connection to efficiently estimate the overlap between the synthesized state and the target state. The protocol involves preparing a 'verification state' by applying a sequence of random single-qubit measurements to the synthesized state, and then measuring the overlap of this verification state with the target state.

The researchers prove that the number of measurements required for accurate certification scales polynomially with the number of qubits, in contrast to existing rigorous protocols that require either deep quantum circuits or exponentially many measurements.

The applications of this new certification technique include:

• Benchmarking quantum systems by verifying the prepared states match the target
• Optimizing quantum circuits to generate desired target states
• Learning and verifying neural network, tensor network, and other representations of quantum states using only single-qubit measurements
• Efficiently predicting hard-to-measure quantum properties from the verified state representations

The researchers demonstrate these applications in numerical experiments with up to 120 qubits, showing advantages over existing methods like cross-entropy benchmarking.

Critical Analysis

The paper presents an innovative and theoretically well-grounded approach to the important problem of efficiently certifying quantum states. The key technical insight of relating certification to random walk mixing times is clever and allows the researchers to sidestep the limitations of previous rigorous protocols.

That said, the paper does not address some potential caveats and limitations of the approach. For example, the analysis assumes that the target state is "almost all" n-qubit states, but it's not clear how restrictive this assumption is in practice. Additionally, the polynomial scaling of measurements with qubits, while an improvement over exponential scaling, may still be challenging for large quantum systems.

Further, the numerical demonstrations, while promising, are still relatively small scale. It would be valuable to see the technique applied to larger, more complex quantum systems to better understand its real-world performance and limitations.

Overall, this is an interesting and technically solid piece of research that offers a new tool for an important problem in quantum information science. But as with any research, there are still open questions and areas for further investigation.

Conclusion

This paper presents a new technique for efficiently certifying that a quantum state synthesized in a lab is close to a target state, using only a small number of single-qubit measurements. This is an important task in quantum information science, but existing rigorous protocols either require complex quantum circuits or exponentially many measurements.

The key innovation is relating the certification problem to the mixing time of a random walk, which allows the researchers to prove that almost all n-qubit target states can be certified from just O(n^2) measurements. This has applications for benchmarking quantum systems, optimizing quantum circuits, and verifying representations of quantum states like neural networks.

While the paper doesn't address all potential limitations, it offers a promising new approach that could significantly improve the practicality of rigorously certifying quantum states, a crucial step in advancing quantum technology. Further research is needed to fully understand the strengths and weaknesses of this technique across a wide range of quantum systems and applications.