Description
Neural Quantum States have emerged as a powerful variational ansatz for modeling complex quantum systems. However, extending this framework to time evolution presents notable challenges. In this work, we simulate the time evolution of quantum states subjected to potential shifts and momentum kicks using a real-valued network architecture governed by the Dirac-Frenkel variational principle. The results are validated by benchmarking wavefunction fidelity and certain physical observables. Ultimately, this work demonstrates the viability of achieving accurate time evolution through machine learning, laying the groundwork for simulating the dynamics of more complex, interacting systems.
References:
A. Romero-Ros, J. Rozalén and A. Rios. Quantum Dynamics with Time-Dependent Neural Quantum States, arXiv 2509.24865, 2025.
L. Hackl, T. Guaita, T. Shi, J. Haegeman, E. Demler and J. I. Cirac. Geometry of variational methods: dynamics of closed quantum systems, SciPost Phys., 9:048, 2020.
A. Lovato, G. Carleo, B. fore, M. Hjorth-Jensen, J. Kim, A. Rios and N. Rocco. Neural-network quantum states for the nuclear many-body problem, arXiv 2602.13826, 2026.