PUBLICATION: HyperMM : Robust multimodal learning with varying-sized inputs © Springer. Personal use of this material is permitted. The definitive version of this paper was published in and is available at : Document