Feasibility of Diagrammatic Monte-Carlo based on weak-coupling expansion in asymptotically free theories: case study of $O(N)$ sigma-model in the large-$N$ limit

TL;DR

This paper explores the potential of Diagrammatic Monte-Carlo methods to compute the mass gap in asymptotically free theories, demonstrating feasibility through the large-$N$ $O(N)$ sigma-model with a convergent double-series expansion.

Contribution

It introduces a double-series expansion for the mass gap in the $O(N)$ sigma-model and shows that Monte-Carlo sampling can effectively compute its coefficients, reducing the sign problem at small coupling.

Findings

Double-series expansion converges quickly to the exact mass gap.

Monte-Carlo sampling of Feynman diagrams is feasible for coefficient calculation.

Sign problem diminishes as coupling approaches the continuum limit.

Abstract

We discuss the feasibility of applying Diagrammatic Monte-Carlo algorithms to the weak-coupling expansions of asymptotically free quantum field theories, taking the large-$N$ limit of the $O(N)$ sigma-model as the simplest example where exact results are available. We use stereographic mapping from the sphere to the real plane to set up the perturbation theory, which results in a small bare mass term proportional to the coupling $\lambda$. Counting the powers of coupling associated with higher-order interaction vertices, we arrive at the double-series representation for the dynamically generated mass gap in powers of both $\lambda$ and $\log(\lambda)$, which converges quite quickly to the exact non-perturbative answer. We also demonstrate that it is feasible to obtain the coefficients of these double series by a Monte-Carlo sampling in the space of Feynman diagrams. In particular, the sign problem of such sampling becomes milder at small $\lambda$, that is, close to the continuum limit.