Dirac Measure on a Group

134 Views Asked by At

In certain works https://arxiv.org/pdf/1803.11173.pdf (page 6, equation 8), the authors use the Dirac measure on the unitary group $\delta(V-U)dU$ to mean that "U must be of form V", where $dU$ is the Haar measure on the unitary group. But the subtraction here bothers me, as this is not the group operation. This would center the measure around the 0 matrix, which is not a unitary. Should this not be written as $\delta(U^*V)dU$ where the Dirac measure is centered around the identity?