Title:Efficient Chromatic-Number-Based Multi-Qubit Decoherence and Crosstalk Suppression

Abstract:The performance of quantum computers is hindered by decoherence and crosstalk, which cause errors and limit the ability to perform long computations. Dynamical decoupling is a technique that alleviates these issues by applying carefully timed pulses to individual qubits, effectively suppressing unwanted interactions. However, as quantum devices grow in size, it becomes increasingly important to minimize the time required to implement dynamical decoupling across the entire system. Here, we present "Chromatic-Hadamard Dynamical Decoupling" (CHaDD), an approach that efficiently schedules dynamical decoupling pulses for quantum devices with arbitrary qubit connectivity. By leveraging Hadamard matrices, CHaDD achieves a circuit depth that scales quadratically with the chromatic number of the connectivity graph for general two-qubit interactions, assuming instantaneous pulses. For the common case of ZZ crosstalk, which is prevalent in superconducting qubit devices, the scaling improves to linear. This represents an exponential improvement over all previous multi-qubit decoupling schemes for devices with connectivity graphs whose chromatic number grows at most polylogarithmically with the number of qubits. For graphs with a constant chromatic number, CHaDD's scaling is independent of the number of qubits. Our results suggest that CHaDD can become a useful tool for enhancing the performance and scalability of quantum computers by efficiently suppressing decoherence and crosstalk across large qubit arrays.