Masked Multi-time Diffusion for Multi-modal Generative Modeling accepted at the NeurIPS 2023 Workshop on Diffusion Models. paper, poster