Taylor coefficients of false Jacobi forms and ranks of unimodal sequences (Q6103486)

Theta functions play a distinguished role in the theory of modular forms, being one of two primary ways to produce modular objects (via Poisson summation; the other, Poincaré series, being a method of averaging). In the past decades, a huge role has been played by variants of theta functions where Poisson summation doesn't quite apply directly. In particular, indefinite theta series have played a central role in the theory of mock modular forms and harmonic Maass since the seminal PhD thesis of Zwegers. These involve quadratic forms which when summed over a full lattice would diverge, so only a piece of a lattice is summed over to obtain convergence. This may be fixed by considering such series as limiting factors of sums with non-holomorhpic prefactors inserted which ``fix'' the Poisson summation method. Other types of ``near miss'' theta functions are the false theta functions and partial theta functions. These two closely related examples are ones where some of the signs in the character are ``wrong'' (false) or only half of a lattice is summed over (partial; often arises in particular when the full lattice sum would vanish due to signs). Partial and false theta functions have until recently had less understood modularity properties, although there were hints of connection with mock modular forms through radial limits of mock modular forms and the theory of quantum modular forms. Recently, \textit{K. Bringmann} and \textit{C. Nazaroglu} [Res. Math. Sci. 6, No. 3, Paper No. 30, 23 p. (2019; Zbl 1456.11074)] developed a framework for modularity of false theta functions. False theta functions natural arise as generating functions in a host of problems in combinatorics, in particular in partition theory. In this paper, the authors use the new modular structure for false theta functions of Bringmann and Nazaroglu to give asymptotics for Taylor coefficients of so-called false Jacobi forms. They use this to produce asymptotic series for rank moments of two types of unimodal sequences. The proofs use the circle method, which is very delicate in this situation. This paper provides a template for future work in applying the Bringmann-Nazaroglu framework to other similar \(q\)-series asymptotics via the Circle Method.