Squarefree values of random polynomials are Gaussian mixtures
In the authors' words
Fix and choose a polynomial uniformly at random from the set of integer polynomials of degree whose coefficients have absolute value at most . We study , the error in the count of squarefree values for . In the range , we prove that converges to a Gaussian mixture that is not itself Gaussian, , where is the weak limit, as , of the prime-indexed vector , is a nonnegative function and is an independent standard Gaussian. This answers a question of Sofos. Finally, in the appendix, we establish that as .
Main resultLimitation the authors admit
Appeared: Thursday, October 22. arXiv. Preprint, not yet peer-reviewed.
Authors' comment: 62 pages, including the appendix