1 paper
Andreas Makris, Paul Fearnhead, Chris Nemeth
Training-free conditional diffusion provides a flexible alternative to task-specific conditional model training, but existing samplers often allocate computation inefficiently: ind…