Exploring quench dynamics as a shortcut to adiabaticity in programmable atoms arrays

The ability to prepare ground states of quantum Hamiltonians via an adiabatic protocol is typically determined by the smallest energy gap during quantum evolution. This poses a challenge for large quantum systems, in particular in instances where the minimum gap scales super-exponentially with system size. We experimentally investigate the breakdown of the quantum adiabatic algorithms for such hard instances of the maximum independent set problem and demonstrate a method to circumvent this limitation. Using QuEra's Aquila programmable quantum simulator based on Rydberg atom arrays, we experimentally realize a hybrid adiabatic-quench-adiabatic protocol as a remedy to the diverging adiabatic timescale and find that it significantly outperforms adiabatic algorithms. We observe quantum-scar-like dynamics for different quench durations, demonstrating that a sweep-quench-sweep approach quantum algorithm can provide a shortcut to adiabaticity for a certain class of problems where adiabatic algorithms fail.