arXiv:2610.05953
Fast mixing of the SYK model at high temperatures
Yiyi Cai, Yongtao Zhan, Alexander Zlokapa
Abstract
The Sachdev-Ye-Kitaev (SYK) model is a strongly interacting fermionic system. Its dynamics are difficult to analyze with Lieb-Robinson bounds and cluster expansions due to its all-to-all disordered interactions. We prove that, with high probability over the disorder, the SYK model admits a quantum Gibbs sampler with a system-size-independent spectral gap at sufficiently high constant temperatures. This yields polynomial-time preparation of SYK Gibbs states on a quantum computer. While prior non-rigorous computations that suggested SYK thermalization were limited to studying local correlators, we emphasize that our result holds up to arbitrary inverse polynomial trace distance. Our proof introduces the pseudo-Lindbladian approach to fermions then applies two key new ideas: a Bochner inequality for quantum dynamics, and an especially simple SYK analysis based on a comparison to classical dynamics.
AI Summary
The paper asks whether SYK Gibbs states can be prepared efficiently with global trace-distance guarantees, despite dense random interactions that obstruct existing high-temperature mixing arguments. It studies an efficiently implementable, KMS-detailed-balanced quantum Gibbs sampler for the quartic Majorana SYK Hamiltonian, with coupling variance $6J^2/N^3$. For $0\leq\beta J\leq10^{-3}$, it proves that, with probability at least $1-2e^{-N}$ over disorder, the sampler has spectral gap at least $507/392>1$ and mixing time $O(N+\log(1/\epsilon))$ in trace norm, uniformly over initial states. These bounds use a generator normalized as a sum of $N$ dissipative terms, without a $1/N$ prefactor. Combined with existing implementation algorithms, this gives polynomial-time quantum preparation to inverse-polynomial trace-distance error. The proof extends pseudo-Lindbladian comparisons to fermions. A Clifford-derivative version of the integrated Bochner argument reduces an auxiliary spectral-gap bound to controlling the Jacobian of imaginary-time-evolved Majoranas. Disorder-averaged moments of its Taylor coefficients are bounded by those of an auxiliary classical quartic spin-glass gradient flow; Gaussian concentration and polynomial derivative estimates then give operator-norm bounds. Related estimates control Gaussian filtering and transfer the gap to the implementable sampler. The guarantee applies only at sufficiently high constant temperature, not throughout SYK’s temperature range, and concerns a designed sampler rather than intrinsic Hamiltonian thermalization. Mixing of order $O(\log N+\log(1/\epsilon))$ remains open. The paper also leaves open quantum advantage, possible classical representations or dequantization, and extensions to other strongly interacting fermionic models.
The main idea behind the proven theorem statement was proposed by ChatGPT 6.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 34 pages; v2 improves exposition and updates references