{"ID":2894753,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2507.10072","arxiv_id":"2507.10072","title":"Frequency Regulation for Exposure Bias Mitigation in Diffusion Models","abstract":"Diffusion models exhibit impressive generative capabilities but are significantly impacted by exposure bias. In this paper, we make a key observation: the energy of predicted noisy samples in the reverse process continuously declines compared to perturbed samples in the forward process. Building on this, we identify two important findings: 1) The reduction in energy follows distinct patterns in the low-frequency and high-frequency subbands; 2) The subband energy of reverse-process reconstructed samples is consistently lower than that of forward-process ones, and both are lower than the original data samples. Based on the first finding, we introduce a dynamic frequency regulation mechanism utilizing wavelet transforms, which separately adjusts the low- and high-frequency subbands. Leveraging the second insight, we derive the rigorous mathematical form of exposure bias. It is worth noting that, our method is training-free and plug-and-play, significantly improving the generative quality of various diffusion models and frameworks with negligible computational cost. The source code is available at https://github.com/kunzhan/wpp.","short_abstract":"Diffusion models exhibit impressive generative capabilities but are significantly impacted by exposure bias. In this paper, we make a key observation: the energy of predicted noisy samples in the reverse process continuously declines compared to perturbed samples in the forward process. Building on this, we identify tw...","url_abs":"https://arxiv.org/abs/2507.10072","url_pdf":"https://arxiv.org/pdf/2507.10072v3","authors":"[\"Meng Yu\",\"Kun Zhan\"]","published":"2025-07-14T08:58:38Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Diffusion Model\"]","has_code":false,"code_links":[{"ID":612129,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2894753,"paper_url":"https://arxiv.org/abs/2507.10072","paper_title":"Frequency Regulation for Exposure Bias Mitigation in Diffusion Models","repo_url":"https://github.com/kunzhan/wpp","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}