Learning holographic horizons

We apply machine learning to understand fundamental aspects of holographic duality, specifically the entropies obtained from the apparent and event horizon areas. We show that simple features of only the time series of the pressure anisotropy, namely the values and half-widths of the maxima and minima, the times these are attained, and the times of the first zeroes can predict the areas of the apparent and event horizons in the dual bulk geometry at all times with a fixed maximum length of the input vector. Given that simple Vaidya-type metrics constructed just from the apparent and event horizon areas can be used to approximately obtain unequal time correlation functions, we argue that the corresponding entropy functions are the information that can be learnt from a single one-point function (without explicit knowledge of the gravitational dynamics) and that is necessary for reconstructing specific aspects of the dual quantum state to the best possible approximations.