only one whose Walsh spectrum and, in particular, nonlinearity, are unknown. This problem
has already been open for 20 years without any progress since the seminal work of
Canteaut, Charpin and Dobbertin from 2000, in which they proved that all Walsh coefficients
of the Dobbertin function over F25m are divisible by 22m. In this paper, we present a
conjecture fully describing the Walsh spectrum of the Dobbertin function. We also show that …