Trotter error and gate complexity of the SYK and sparse SYK models (2502.18420v1)
Abstract: The Sachdev-Ye-Kitaev (SYK) model is a prominent model of strongly interacting fermions that serves as a toy model of quantum gravity and black hole physics. In this work, we study the Trotter error and gate complexity of the quantum simulation of the SYK model using Lie-Trotter-Suzuki formulas. Building on recent results by Chen and Brandao (arXiv:2111.05324), we derive bounds on the first- and higher-order Trotter error of the SYK model, and subsequently find near-optimal gate complexities for simulating these models using Lie-Trotter-Suzuki formulas. For the $k$-local SYK model on $n$ Majorana fermions, our gate complexity estimates for the first-order Lie-Trotter-Suzuki formula scales with $O(n{k+\frac{5}{2}}t2)$ for even $k$ and $O(n{k+3}t2)$ for odd $k$, and the gate complexity of simulations using higher-order formulas scales with $O(n{k+\frac{1}{2}}t)$ for even $k$ and $O(n{k+1}t)$ for odd $k$. Given that the SYK model has $\Theta(nk)$ terms, these estimates are close to optimal. These gate complexities can be further improved when simulating the time-evolution of an arbitrary fixed input state $|\psi\rangle$, leading to a $O(n2)$-reduction in gate complexity for first-order formulas and $O(\sqrt{n})$-reduction for higher-order formulas. We also apply our techniques to the sparse SYK model, a simplified variant of the SYK model obtained by deleting all but a $\Theta(n)$ fraction of the terms in a uniformly i.i.d. manner. We compute the average (over the random term removal) gate complexity for simulating this model using higher-order formulas to be $O(n2 t)$, a bound that also holds for a general class of sparse Gaussian random Hamiltonians. Similar to the full SYK model, we obtain a $O(\sqrt{n})$-reduction simulating the time-evolution of an arbitrary fixed input state $|\psi\rangle$.
Collections
Sign up for free to add this paper to one or more collections.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.