Insights from the Quantum Era - June 2024

An Abstraction Hierarchy Toward Productive Quantum Programming

This insightful viewpoint paper delves into abstraction hierarchies for quantum programming, drawing inspiration from parallel computing. It proposes a three-level hierarchy encompassing programming, execution, and hardware layers. While the abstraction is not flawless—particularly for near-term devices where hardware constraints can impact all levels—it nonetheless encourages a broader perspective on representing and compiling quantum and hybrid programs beyond the circuit level. This broader perspective is crucial for optimizing hardware performance. 

Read here on arXiv

‍

Analog Counterdiabatic Quantum Computing

This work utilized counterdiabatic analog quantum protocols to solve graph optimization problems on neutral atom quantum hardware. Counterdiabatic protocols suppress transitions away from the ground state during an adiabatic process, thereby enhancing the probability of finding solutions to optimization problems. The variationally optimized counterdiabatic protocols were benchmarked on both QuEra’s Aquila and Pasqal’s Fresnel analog quantum processors, marking the first such comparison. The results clearly demonstrate the benefits of counterdiabatic driving. 

read here on arXiv

‍

{{Newsletter-logo}}

‍

Recurrent neural network wave functions for Rydberg atom arrays on kagome lattice

Generating highly entangled ground states of various systems is a key application of analog computers like Aquila. Notably, this includes topological spin liquids, which are difficult to simulate using classical methods such as the density matrix renormalization group. In this study, collaborators from Canada, the USA, and Switzerland introduce an innovative approach by integrating machine learning techniques. Using two-dimensional recurrent neural network (RNN) wave functions, they explore the ground states of Rydberg atom arrays on the Kagome lattice. Their findings reveal that in the highly frustrated and highly entangled regimes, the RNN predicts a paramagnetic phase and the absence of a spin-glass phase. 

Read here on arXiv

‍

Prediction of chaotic dynamics and extreme events: A recurrence-free quantum reservoir computing approach

Quantum reservoir computing is a prominent method for practical quantum machine learning. In this study, researchers from Imperial College London present a novel approach to utilizing quantum reservoirs for time-series prediction, a key application of this methodology. Their simulations demonstrate computational scalability and performance, indicating a potential for longer prediction capacity compared to classical reservoir approaches. 

Read here on arXiv