Machine-learned nodal structures of Fermion systems
Authors:
Preprint
Abstract
A major challenge in quantum physics is the accurate simulation of fermionic systems, particularly those involving strong correlations. While effective for bosonic systems, traditional quantum Monte Carlo methods encounter the notorious sign problem when applied to Fermions, often resulting in biased outcomes through the fixed-node approximation. This work demonstrates the potential of machine learning techniques to address these limitations by allowing nodal structures to be learned through gradient descent optimization iterations and the variational algorithm. Using a neural network to represent the wave function, we focus on quantum dots containing up to 30 electrons. The results show a significant reduction in the variational bias, achieving greater accuracy and a lower ground state energy than diffusion Monte Carlo with the fixed-node approximation. Our approach paves the way for precise and accurate property predictions in fermionic strongly correlated systems, advancing fundamental understanding and applications in quantum technologies.Keywords
All Related Versions
This publication has 46 references indexed in Scilit:
- Low disorder and high valley splitting in siliconnpj Quantum Information, 2024
- Modeling $^4$He${_N}$ Clusters with Wave Functions Based on Neural NetworksJournal of Low Temperature Physics, 2024
- Single-Electron Occupation in Quantum Dot Arrays at Selectable Plunger Gate VoltageNano Letters, 2023
- Simultaneous single-qubit driving of semiconductor spin qubits at the fault-tolerant thresholdNature Communications, 2023
- Experimental realization of an extended Fermi-Hubbard model using a 2D lattice of dopant-based quantum dotsNature Communications, 2022
- A perspective on machine learning and data science for strongly correlated electron problemsCarbon Trends, 2022
- Double Excitation Energies from Quantum Monte Carlo Using State-Specific Energy OptimizationJournal of Chemical Theory and Computation, 2022
- Energy Derivatives in Real-Space Diffusion Monte CarloJournal of Chemical Theory and Computation, 2021
- Semiconductor Quantum Dots for Memories and Neuromorphic Computing SystemsChemical Reviews, 2020
- Variational Monte Carlo for Interacting Electrons in Quantum DotsJournal of Low Temperature Physics, 2005