DARE mitigates neglect of important tokens in conditional diffusion models via distribution-rectified guidance and spatial attention alignment.
As shown in Table 3, after introducing DARE, the performance improves across multi- ple evaluation aspects compared with Seedance (Gao et al., 2025)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Not all tokens contribute equally to diffusion learning
DARE mitigates neglect of important tokens in conditional diffusion models via distribution-rectified guidance and spatial attention alignment.