Structure, Positivity and Classical Simulability of Kirkwood-Dirac Distributions

Abstract: The Kirkwood-Dirac (KD) quasiprobability distribution is known for its role in quantum metrology, thermodynamics, as well as the foundations of quantum mechanics. Here, we study the superoperator evolution of KD distributions and show that unitaries which preserve KD positivity do not always correspond to a stochastic evolution of quasiprobabilities. Conversely, we show that stochastic KD superoperators are always induced by generalised permutations within the KD reference bases. We identify bounds for pure KD positive states in distributions defined on mutually unbiased bases, showing that they always form uniform distributions, in full analogy to the stabilizer states. Subsequently, we show that the discrete Fourier transform of KD distributions on qudits in the Fourier basis follows a self-similarity constraint and provides the expectation values of the state with respect to the Weyl-Heisenberg unitaries, which can then be transformed into the (odd-dimensional) Wigner distribution. This defines a direct mapping between the Wigner, and qudit KD distributions without a reconstruction of the density matrix. Finally, we identify instances where the classical sampling-based simulation algorithm of Pashayan et al. [Phys. Rev. Lett. 115, 070501] becomes exponentially inefficient in spite of the state being KD positive throughout its evolution.