The Sine-beta process is a mathematical object from random matrix theory that describes how eigenvalues of certain random matrices are distributed. A key quantity for understanding this process is the "pair correlation function," which measures the likelihood of finding two eigenvalues at a given distance from each other. The parameter beta acts like an inverse temperature, controlling how strongly eigenvalues repel one another. Earlier work by Qu and Valko developed tools for studying this function but left open some technical gaps that this paper closes.
The first contribution is proving that the pair correlation function varies smoothly in a joint sense when you change both the inverse temperature beta and the spatial separation simultaneously. Previous work had only established this for beta greater than 2. The authors get around the restriction by using a clever general principle: if a family of probability distributions changes continuously in two separate parameters, and one of those parameters has a natural ordering structure (where increasing it makes outcomes systematically larger or smaller), then continuity in both parameters together follows for free. This avoids messy Fourier analysis that previous approaches required.
The second contribution connects two seemingly different mathematical formulas that both describe the same pair correlation function when beta is an even integer. One formula comes from the Qu-Valko recursion, built from matrix operations, and the other comes from classical Selberg integrals, which are important in combinatorics and physics. The paper shows these are actually the same by constructing an explicit mathematical transformation that converts one description into the other. Along the way, this identification gives a concrete meaning to every coefficient in the Qu-Valko recursion: each one equals a specific statistical moment of a probability distribution built from Jacobi polynomials, providing a satisfying dictionary between two previously separate areas of the theory.