We propose a revised definition of quasi-distributions within the framework of large-momentum effective theory (LaMET) that improves convergence towards the large-momentum limit. Since the definition of quasi-distributions is not unique, each choice goes along with a specific matching function, we can use this freedom to optimize convergence towards the large-momentum limit. As an illustration, we study quasi-distributions with a Gaussian weighting factor that naturally suppresses long-range correlations, which are plagued by artifacts. This choice has the advantage that the matching functions can be trivially obtained from the known ones. We apply the Gaussian weighting to the previously published results for the nonperturbatively renormalized unpolarized quark distribution, and find that the unphysical oscillatory behavior is significantly reduced.
Download