RL-Adapt uses single-shot reinforcement learning to optimize ground state preparation via dissipative dynamics. Achieves 7.6× faster convergence on spin and electronic structure systems without requiring a priori knowledge of system spectrum.
@informaq.bsky.social