Diffusion models can extract reusable density-mode concepts from their time-indexed scores to enable compositional generation at test time on held-out benchmarks from ColorMNIST and CelebA.
A phase transition in diffusion models reveals the hierarchical nature of data.Proceedings of the National Academy of Sciences, 122(1):e2408799121
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
DAR replaces residual addition in DiTs with learnable, timestep-adaptive aggregation of sublayer outputs, yielding 2.11 FID improvement on SiT-XL/2 and 8.75x faster convergence on ImageNet 256x256.
citing papers explorer
-
Test-Time Compositional Generalization in Diffusion Models via Concept Discovery
Diffusion models can extract reusable density-mode concepts from their time-indexed scores to enable compositional generation at test time on held-out benchmarks from ColorMNIST and CelebA.
-
Rethinking Cross-Layer Information Routing in Diffusion Transformers
DAR replaces residual addition in DiTs with learnable, timestep-adaptive aggregation of sublayer outputs, yielding 2.11 FID improvement on SiT-XL/2 and 8.75x faster convergence on ImageNet 256x256.