Object-Centric Image Generation with Factored Depths, Locations, and Appearances. arxiv.org/abs/2004.00642

Some random musings on how the whole "AI scaling" debate. It is clear that neural nets can model any predictive distribution p(x_future|x_past) given enough parameters and data, since they are universal approximators;

Happy to announce our work on enhancing out-of-distribution generalisation in object-centric world models has been accepted at #NeurIPS2022! A quick summary... UofG Computing Science UoG IDA Section, Glasgow CS School of Informatics, The University of Edinburgh Snap Inc. NeurIPS Conference

If you're interested in how image diffusion can be adapted to learn about the 3D world, don't miss our poster (25) presentation this afternoon at #CVPR2023.