Error mitigation has been considered by many to be the salvation of noisy quantum computers. In this work, collaborators from Caltech, Harvard, and JILA demonstrate, among other results, that quantum circuits for which error mitigation is efficient are classically simulable. However, the noise biasing and high gate accuracies introduced by error-correction schemes render fault-tolerant circuits hard to simulate classically. While accuracy constraints may lead some classical simulations of NISQ circuits to be quite extensive, this work strongly establishes that error correction is indeed the most promising path forward to quantum advantage.
In a collaboration between QuEra and Purdue University, an “adversarial quantum autoencoder model” is introduced to evaluate optimization problems in analog quantum computers, overcoming restrictions due to encoding and constraint overheads. The method relies on quantum-enhanced Boltzmann sampling, and its effectiveness is demonstrated in optimizing the design of photonic devices.
{{Newsletter-signup}}
Quantum LDPC codes have established their reputation for spatial efficiency in encoding logical qubits. Yet, so far, schemes for performing actual computation (gates) with such encoded qubits have often been serialized and constrained, leading to inefficiency in operational time. This work makes significant advances on this front, describing schemes for the selective and parallel addressing of quantum LDPC logical qubits. This brings a space-time efficient implementation of quantum algorithms on quantum LDPC codes closer to reality.
The fidelity of operations with neutral-atom qubits has recently seen considerable advances. To bring this to the next level, a better understanding of error sources and control schemes is necessary. In this work, a Caltech-based team describes a detailed ab-initio model for strontium qubits, designing and demonstrating time-optimal CZ gates with 99.7% fidelity. In this process, the team predicts a pathway for CZ gates with over 99.9% fidelity.
.svg){kind=small}
