AI and the World's Hard Problems · Quantum physics & quantum computing · arXiv · 2026-10-21
The surprising simplicity of forecasting quantum chaotic many-body dynamics
Accurately simulating chaotic quantum many-body dynamics remains a major challenge for classical methods because entanglement builds up and spreads rapidly during evolution. The study casts quantum dynamics as a time-series forecasting problem on a few-qubit PXP chain realizable with Rydberg-atom arrays, comparing an expressive nonlinear Transformer with the simple linear model DLinear across regimes from ergodic behavior to quantum many-body scarring. The Transformer degrades as initial states approach the scarred limit, whereas DLinear stays accurate throughout, showing that observables of complex quantum evolution can be forecast with high accuracy through a simple linear mapping from past to future observations.