We report on a use of low thickness-to-diameter aspect-ratio pores and machine learning for discriminating single-particles. In stark contrast to the conventional resistive pulse analysis that looks only on the height of the ionic current spike signals obtained with long fluidic channels to deduce the particle size, the present technique goes multimodal with state-of-the-art machine learning algorithms to extract and compare multiple features of each electrical signatures in a low-aspect-ratio pore sensor whereby offering a way to identify micro- and nano-objects by their multiplex physical properties.