Spectral Prior for Reducing Exposure Bias in Diffusion Models

• URL: http://arxiv.org/abs/2607.22091v1
• Date: Fri, 24 Jul 2026 08:37:04 GMT
• Title: Spectral Prior for Reducing Exposure Bias in Diffusion Models
• Abstract summary: We propose a lightweight, guidance-based method that calibrates the power spectrum of intermediate predictions to a pre-computed prior.<n>Our approach consists of two stages: (1) offline fitting of a parametric spectrum model from training data, and (2) inference of exposure via efficient FFT-based gradient.
• Score: 37.66268273278838
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Diffusion models typically suffer from error accumulation during iterative sampling, commonly referred to as exposure bias. We reveal systematic frequency-dependent discrepancies between training and inference, which can be interpreted as frequency-dependent SNR error. Crucially, the direction of this mismatch varies across models and timesteps, indicating that fixed correction rules do not generalize. We propose Spectral Alignment (SPA), a lightweight, guidance-based method that calibrates the power spectrum of intermediate predictions to a pre-computed prior. Our approach consists of two stages: (1) offline fitting of a parametric spectrum model from training data, and (2) inference-time guidance via efficient FFT-based gradient computation. SPA introduces minimal computational overhead (3-4\%) and is complementary to Classifier-Free Guidance (CFG). We demonstrate consistent improvements across diverse architectures, from pixel-space models (DDPM, ADM) to latent diffusion models (SD2.0, SDXL) and flow-matching models (SD3.5, FLUX). Our implementation is available at https://github.com/SonyResearch/SPA.

Related papers

• You Only Flow Once: Calibrated and Real-Time Radar Pose Estimation with Multi-Hypothesis Normalizing Flows [5.241472125982733]
  radar point clouds often correspond to multiple plausible human poses, making deterministic pose estimation fundamentally ill-posed.<n> Diffusion-based alternatives can model multi-hypothesis distributions but require costly sequential denoising for each distribution sample.<n>We propose Multi-Hypothesis Normal Flow Generator (MH-G), which models pose distributions from radar point clouds using a conditional flow backbone.
  arXiv  Detail & Related papers  (2026-08-10T13:15:09Z)
• Score Distillation of Flow Matching Models [67.86066177182046]
  We extend Score identity Distillation (SiD) to pretrained text-to-image flow-matching models.<n>SiD works out of the box across these models, in both data-free and data-aided settings.<n>This provides the first systematic evidence that score distillation applies broadly to text-to-image flow matching models.
  arXiv  Detail & Related papers  (2025-09-29T17:45:48Z)
• DS-Diffusion: Data Style-Guided Diffusion Model for Time-Series Generation [3.7098771725459336]
  We propose a data style-guided diffusion model (DS-Diffusion) for time series generation tasks.<n>The DS-Diffusion avoids retraining the entire framework to introduce conditional guidance.<n>The generated samples can clearly indicate the data style from which they originate.
  arXiv  Detail & Related papers  (2025-09-23T03:06:39Z)
• Single-Step Reconstruction-Free Anomaly Detection and Segmentation via Diffusion Models [1.1487074612765584]
  We introduce Reconstruction-free Anomaly Detection with Attention-based diffusion models in Real-time (RADAR)<n>RADAR overcomes the limitations of reconstruction-based anomaly detection.<n>We evaluate RADAR on real-world 3D-printed material and the MVTec-AD dataset.
  arXiv  Detail & Related papers  (2025-08-06T18:56:08Z)
• Efficient and Unbiased Sampling from Boltzmann Distributions via Variance-Tuned Diffusion Models [24.911598225979606]
  Variance-Tuned Diffusion Sampling (VT-DIS) is a lightweight method that adapts the per-step noise covariance of a pretrained score-based diffusion model.<n>VT-DIS achieves effective sample sizes of approximately 80 %, 35 %, and 3.5 %, respectively, on the DW-4, LJ-13, and alanine-dipeptide benchmarks.
  arXiv  Detail & Related papers  (2025-05-27T10:37:48Z)
• GLAD: Towards Better Reconstruction with Global and Local Adaptive Diffusion Models for Unsupervised Anomaly Detection [60.78684630040313]
  Diffusion models tend to reconstruct normal counterparts of test images with certain noises added. From the global perspective, the difficulty of reconstructing images with different anomalies is uneven. We propose a global and local adaptive diffusion model (abbreviated to GLAD) for unsupervised anomaly detection.
  arXiv  Detail & Related papers  (2024-06-11T17:27:23Z)
• One More Step: A Versatile Plug-and-Play Module for Rectifying Diffusion Schedule Flaws and Enhancing Low-Frequency Controls [77.42510898755037]
  One More Step (OMS) is a compact network that incorporates an additional simple yet effective step during inference. OMS elevates image fidelity and harmonizes the dichotomy between training and inference, while preserving original model parameters. Once trained, various pre-trained diffusion models with the same latent domain can share the same OMS module.
  arXiv  Detail & Related papers  (2023-11-27T12:02:42Z)
• Diffusion models for probabilistic programming [56.47577824219207]
  Diffusion Model Variational Inference (DMVI) is a novel method for automated approximate inference in probabilistic programming languages (PPLs) DMVI is easy to implement, allows hassle-free inference in PPLs without the drawbacks of, e.g., variational inference using normalizing flows, and does not make any constraints on the underlying neural network model.
  arXiv  Detail & Related papers  (2023-11-01T12:17:05Z)
• Elucidating the Exposure Bias in Diffusion Models [7.529885833916501]
  We investigate the input mismatch between training and sampling in diffusion models. We propose a training-free method called Epsilon Scaling to alleviate the exposure bias. We show that Epsilon Scaling explicitly moves the sampling trajectory closer to the vector field learned in the training phase.
  arXiv  Detail & Related papers  (2023-08-29T14:16:09Z)
• Reflected Diffusion Models [93.26107023470979]
  We present Reflected Diffusion Models, which reverse a reflected differential equation evolving on the support of the data. Our approach learns the score function through a generalized score matching loss and extends key components of standard diffusion models.
  arXiv  Detail & Related papers  (2023-04-10T17:54:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.