Selected Configuration Interaction Using Time-Evolved Population Statistics

Area of Science:

• Quantum computing applications in chemistry
• Molecular electronic structure theory
• Computational quantum chemistry

Background:

• Selected Configuration Interaction (SCI) constructs adaptable configuration subspaces for accurate molecular electronic structure calculations.
• Quantum-SCI (QSCI) leverages quantum resources to guide subspace construction, addressing classical computational challenges in state sampling and dynamics.
• Existing QSCI methods often rely on error-mitigated quantum measurements to form subspaces.

Purpose of the Study:

• To propose an alternative QSCI approach that integrates classically inaccessible quantum information into the selection criteria.
• To enhance subspace expansion pathways by biasing selection using quantum state population statistics.
• To improve the accuracy and compactness of wave function descriptions in electronic structure calculations.

Main Methods:

• Developed an occupancy-guided expansion method using population statistics of time-evolved quantum states to predict important configurations.
• Combined quantum-guided subspace expansion with direct sampling of configurations from the quantum state.
• Incorporated multireference perturbation theory to account for correlations outside the selected subspace.
• Implemented and demonstrated the approach on a 42-qubit IQM superconducting device for the SiH4 molecule.

Main Results:

• The occupancy-guided expansion successfully biases subspace selection, complementing direct sampling and allowing higher-order excitations.
• The quantum-SCI method was successfully applied to compute the potential energy curve of SiH4.
• Benchmarking against Heatbath CI assessed the compactness and accuracy of the resulting wave function.

Conclusions:

• The proposed occupancy-guided QSCI approach offers a novel pathway for constructing accurate configuration interaction wave functions.
• This method effectively injects quantum-derived information into the selection process, potentially overcoming classical limitations.
• The successful hardware demonstration highlights the potential of quantum computing for advancing molecular electronic structure theory.