Title:Controllable Incremental Algorithm for Entanglement Entropy and Other Observables with Exponential Variance Explosion in Many-Body Systems

Abstract:Researchers in the field of physical science are continuously searching for universal features in strongly interacting many-body systems. However, these features can often be concealed within highly complex observables, such as entanglement entropy (EE). The non-local nature of these observables makes them challenging to measure experimentally or evaluate numerically. Therefore, it is of utmost importance to develop a reliable and convenient algorithm that can accurately obtain these complex observables. In this paper, with help of quantum Monte Carlo (QMC), we reveal that the statistical variance of EE exponentially explodes with respect to the system size, making the evaluation of EE inaccurate. We further introduce an incremental algorithm based on the framework of QMC to solve this conundrum. The total number of our incremental processes can be quantitatively determined and reasonably adjusted, making it easy to control the precision in practice. We demonstrate the effectiveness and convenience of our incremental algorithm by using it to obtain the highly accurate EE of a 2D Hubbard model as an example. Additionally, our algorithm can be potentially generalized to calculate other numerically statistically unstable observables with exponential variance growth, such as the entanglement spectrum and topological entanglement negativity of correlated boson/spin and fermion systems, as well as other general functions of determinants of Green's functions in interacting fermions. Accurately measuring these complex observables has the potential to inspire the development of physical theories and guide the direction of experiments.