Global Poles of the Two-Loop Six-Point N=4 SYM integrand
Kasper J. Larsen
TL;DR
This work validates that the leading-singularity method in strictly four dimensions reproduces the full two-loop six-point MHV integrand of $\mathcal{N}=4$ SYM as predicted by recent loop-level recursion relations. By computing maximal and heptacuts of both general and factorized double-box topologies and solving the resulting linear systems, the parity-even and parity-odd components of the integrand are shown to match numerically with the recursion-based predictions across many random momenta. The analysis also clarifies the role of multidimensional contours and the overcomplete basis in determining integral coefficients, and discusses the practical utility of the intermediate results for constructing maximal-cut contours once complete IBP relations are known. Overall, the paper strengthens confidence in the consistency between modern four-dimensional recursion approaches and traditional unitarity-based methods for two-loop six-point amplitudes, while outlining avenues for future extensions to NMHV sectors and more extensive IBP-driven contour determinations.
Abstract
Recently, a recursion relation has been developed, generating the four-dimensional integrand of the amplitudes of N=4 supersymmetric Yang-Mills theory for any number of loops and legs. In this paper, I provide a comparison of the prediction for the two-loop six-point maximally helicity-violating (MHV) integrand against the result obtained by use of the leading singularity method. The comparison is performed numerically for a large number of randomly selected momenta and in all cases finds agreement between the two results to high numerical accuracy.
