RealFeedApp Using NQS as a trial state eliminates the bottleneck of summing matchgate shadows
Researchers replaced the computational bottleneck in the QC-AFQMC algorithm with a symmetry-constrained neural quantum state, successfully recovering the ideal-upCCD barrier and reaction energy for a nickel-catalyzed oxidative-addition reaction within statistical uncertainty.
- Field
- tools and releases
- What they did
- Researchers developed a neural quantum state model that replaces heavy shadow-based computations with fast predictions of energies and overlaps in the quantum Monte Carlo algorithm.
- Why it matters
- This enables faster calculations for chemical reactions, such as nickel-catalyzed oxidative addition, yielding accurate results without the need to average data across thousands of snapshots.
#quantum computing#quantum monte carlo#neural quantum states#quantum chemistry#transformer
