5 June 2025 Thu 4 pm

Quantum computation is rapidly evolving, not only as a powerful tool to simulate and understand complex materials but also as a novel synthetic quantum system in itself, exhibiting rich, universal physical phenomena. This talk will explore both these facets, drawing from our recent work.

First, we delve into the application of quantum machine learning (QML) for materials science, specifically for the classification of quantum phases of matter. We introduce a channel attention mechanism for Quantum Convolutional Neural Networks (QCNNs). This mechanism significantly improves the predictive accuracy of QCNNs in identifying distinct quantum phases, such as those found in the cluster Ising model. Our approach innovatively utilizes measurements on qubits, typically discarded in conventional QCNN models, to create multiple, weighted channels of the output state. This not only boosts performance but also proves more efficient, outperforming standard hybrid QCNN-classical neural network strategies with a markedly smaller number of parameters.

Second, we shift our focus to the intrinsic behavior of quantum computations, treating them as complex dynamical systems. We investigate monitored quantum dynamics in systems of free fermions, deriving universal Fokker-Planck equations that precisely describe the stochastic time evolution of the entire density-matrix spectra. Remarkably, these equations bear a strong resemblance to those governing transmission eigenvalue statistics in mesoscopic disordered systems. We identify universal fluctuations of Rényi entropies in the chaotic regime of monitored dynamics. These fluctuations are independent of microscopic details, serving as a nonunitary analogue of the universal conductance fluctuations (UCF) in mesoscopic metals, and their specific values depend on the fundamental symmetry class of the dynamics.

References:

[1] G. Budiutama, et al., "Channel attention for quantum convolutional neural networks," Phys. Rev. A 110, 012447 (2024).

[2] Z. Xiao, T. Ohtsuki, and K. Kawabata, "Universal Stochastic Equations of Monitored Quantum Dynamics," Phys. Rev. Lett. 134, 140401 (2025).

< Previous

Next >