Joonho Lee [Harvard]

Abstract: In this talk, I will present two quantum-classical hybrid algorithms for computational chemistry that my group is actively developing. The first builds on our previous work on quantum classical hybrid quantum Monte Carlo (QC-QMC), which leverages classical shadow techniques to improve the accuracy of ground-state QMC simulations. I will discuss the key challenges encountered when scaling this approach to larger systems and the strategies we have developed to address them. The second algorithm uses quantum simulators to extract memory kernels that enable efficient prediction of future dynamics on a classical computer. I will present a detailed error analysis of this approach and discuss how its performance scales with noise. Together, these results highlight concrete pathways toward achieving practical quantum advantage in chemically relevant problems.