cs.LGOct 3, 2025

DELTA: Robustly Training Label-Conditional Diffusion Models with Weak Annotations

Authors: Dong-Dong Wu, Jiacheng Cui, Wei Wang, Zhiqiang Shen, Masashi Sugiyama

Organizations: The University of Tokyo, Tokyo, Japan · RIKEN AIP, Tokyo, Japan · Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE

Abstract

While label-conditional diffusion models have recently exhibited remarkable generative capabilities, their success heavily relies on massive, cleanly labeled datasets. In practice, categorical supervision is rarely perfect: it is often corrupted by noise, clouded by ambiguity, or partially missing. Training directly on such weak annotations severely degrades generation quality, and existing robust methods offer only fragmented solutions that depend on scarce auxiliary priors. To overcome these limitations, we propose DELTA, a unified framework that robustly trains Diffusion modELs across all three weak annoTAtion types. By treating the unknown true label as a latent variable, DELTA optimizes a principled variational objective that jointly recovers the clean data distribution and infers the true label posteriors. To make this joint training computationally tractable, we further introduce a median-centered timestep sampling strategy that efficiently concentrates estimation where the diffusion process is most informative. Extensive experiments on synthetic and real-world benchmarks demonstrate that DELTA produces high-fidelity, class-consistent samples across diverse weak supervision scenarios, outperforming specialized baselines while demanding strictly less prior knowledge.

Figures & tables

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Semi-Supervised Conditional Diffusion via Label Augmentation

    Jul 18, 2026Jin Su, Yuan Gao, Yong Zhou +1Data AugmentationSemi-Supervised Learning

  2. LACON: Training Text-to-Image Model from Uncurated Data

    Mar 27, 2026Zhiyang Liang, Ziyu Wan, Hongyu Liu +4Geometry-Conditioned Image GenerationT2I Generation