Orateur
Curt von Keyserlingk
(KCL)
Description
Can we efficiently estimate transport coefficients (conductivities etc) in many-body quantum systems using classical computers? Drawing on lessons learned from studying scrambling and entanglement entropy dynamics in generic many-body systems, I propose an upper bound on the computational resources required to simulate transport at high temperatures: CPU time/memory $\sim e^{O(\log(\epsilon^{-1}))^2}$, where $\epsilon$ is the desired degree of precision. I'll describe a method (DAOE) achieving this bound. I'll explain why DAOE in its original form fails at low temperatures/extreme filling, and then propose an extension of it which shows encouraging signs of working in these limits.
