871 papers
This paper proposes a two-stage neural residual decoding framework that combines a pretrained graph neural network for local confidence-based pre-decoding with a global NVIDIA-accelerated minimum-weight perfect matching cleanup, achieving significant improvements in both logical failure rates and decoding speed across various code distances and noise models compared to traditional methods.
This first-principles study investigates the pressure-induced metallization and electron-phonon-mediated superconductivity in compressed alkali bromides and iodides, revealing that the superconducting transition temperature is primarily governed by electron-phonon coupling strength and is more closely linked to the ground-state crystal structure than to high-pressure phases.
This study predicts that electrons on superfluid helium exhibit a large, positive motional anharmonicity that enables high-fidelity gates and strong-dispersive readout, thereby identifying coherence times (s and s) as the primary remaining barrier to realizing a viable qubit platform.
This paper applies the Hirota bilinear method for the first time to derive explicit single, double, triple, and general -soliton solutions for the Qiao-Li equation, while numerical simulations reveal its unique valley-peak transformation dynamics during soliton collisions.
This study demonstrates that in sputter-grown amorphous FeCuB films, the epsilon near zero (ENZ) condition coincides with the disorder-driven metal-insulator transition on the metallic side, as the dielectric function approaches zero at metal concentrations similar to those near the transition point.
This paper proposes an updated brilliance characterization for advanced accelerator-based x-ray sources using a 0.01% bandwidth metric to correct overestimations in longitudinally coherent regimes, while providing more accurate general expressions and a closed-form saturation length by accounting for non-uniform electron profiles and pulse duration scaling.
This study employs first-principles DFT calculations to comprehensively characterize the structural, electronic, optical, thermodynamic, and vibrational properties of various monolayer and bilayer SnTe phases, identifying the γ-bilayer and π-monolayer as dynamically stable candidates with promising potential for spintronic, optoelectronic, and thermoelectric applications.
This study employs a Physics-Informed Machine Learning framework to systematically screen approximately 1,000 lead-free Sn–Ge perovskite compositions, successfully identifying a robust stability window that balances structural integrity with optimal bandgap properties for next-generation, non-toxic photovoltaic applications.
This paper introduces a robust quantum-illumination protocol using two receivers and a dual-homodyne strategy to achieve stable, high-precision spatial positioning beyond the far-field limit that universally saturates the quantum Cramér–Rao bound and remains insensitive to parameter fluctuations in noisy environments.
This study validates the 11B(p,n)11C reaction for producing high-purity Carbon-11 using natural boron targets by integrating TALYS simulations with TR-19 cyclotron experiments, demonstrating that boron nitride yields the highest activity (12.37 GBq) and confirming the method's suitability for future laser-driven production at ELI-NP.