August 28th, 2024

High-Temperature Gibbs States Are Unentangled and Efficiently Preparable

The paper reveals that Gibbs states of local Hamiltonians become separable above a certain temperature, allowing efficient sampling and challenging existing beliefs about quantum correlations, with no super-polynomial speedups in preparation.

Read original articleLink Icon
High-Temperature Gibbs States Are Unentangled and Efficiently Preparable

The paper titled "High-Temperature Gibbs States are Unentangled and Efficiently Preparable" by Ainesh Bakshi and colleagues presents significant findings regarding thermal states of local Hamiltonians. The authors demonstrate that above a certain constant temperature, specifically for a local Hamiltonian on a graph with degree \( \mathfrak{d} \), the Gibbs state becomes separable. This means that for inverse temperatures \( \beta < 1/(c\mathfrak{d}) \), where \( c \) is a constant, the Gibbs state can be represented as a classical distribution over product states. This finding challenges the traditional understanding of short-range quantum correlations in Gibbs states. Furthermore, the authors establish that it is possible to efficiently sample from this distribution, allowing for the preparation of a state that is close to the Gibbs state using a depth-one quantum circuit and polynomial classical overhead, specifically for \( \beta < 1/(c\mathfrak{d}^3) \). The results indicate that the preparation of Gibbs states does not offer super-polynomial quantum speedups at temperatures above a fixed constant, which has implications for quantum computing and thermal state manipulation.

- The Gibbs states of local Hamiltonians are separable above a constant temperature.

- Efficient sampling from the distribution over product states is possible.

- The findings challenge existing beliefs about quantum correlations in Gibbs states.

- Preparation of Gibbs states does not yield super-polynomial speedups at high temperatures.

Link Icon 5 comments
By @johntb86 - 8 months
By @empath75 - 8 months
I'm very much approaching this from a level of "I know what some of those words mean", but is this result surprising that quantum systems lose entanglement at high temperatures? Isn't that the whole point of running quantum computers at very low temperatures?
By @0134340 - 8 months
So is this another blow for the free will argument based on quantum entanglement that some posited was happening in the brain? It seems this is another nail in the coffin leaving very few arguments other than mostly solipsistic ones.
By @beefman - 8 months