paper-with-me

홈 › Papers

Elucidating Representation Degradation Problem in Diffusion Model Training

2026-05-11 · Zhipeng Yao, Dazhou Li, Zitong Zhang, Durude Mahee, Fan Zhu, Wenbin Zhang, Xinwei He, Yeying Jin, Rui Yu arxiv

Diffusion models have achieved remarkable success, yet their training remains inefficient due to a severe optimization bottleneck, which we term Representation Degradation. As noise levels increase, the outputs of the trained model exhibit progressive structural distortion, which can destabilize training and impair generation quality. Our analysis suggests that this instability is driven by mismatched target recoverability, which is associated with Neural Tangent Kernel (NTK) spectral weakening and effective low-rank behavior. To address this, we propose Elucidated Representation Diffusion (ERD), a plug-and-play framework that dynamically reallocates optimization effort according to effective recoverability. By stabilizing representation learning without external supervision, ERD accelerates convergence and achieves strong empirical performance across diffusion backbones.

📄 PDF Abstract BibTeX arXiv:2605.10790

Code (0)

등록된 구현이 없습니다.

Tasks

Representation Learning

Similar Papers 제목 키워드 기반

LLDiffusion: Learning Degradation Representations in Diffusion Models for Low-Light Image Enhancement

2023-07-27 · Tao Wang, Kaihao Zhang, Ziqian Shao, Wenhan Luo 외

Current deep learning methods for low-light image enhancement (LLIE) typically rely on pixel-wise mapping learned from paired data. However, these methods often overlook the importance of considering degradation represen…

Image EnhancementImage GenerationLow-Light Image Enhancement

DR-BFR: Degradation Representation with Diffusion Models for Blind Face Restoration

2024-11-15 · Xinmin Qiu, Bonan Li, ZiCheng Zhang, Congying Han 외

Blind face restoration (BFR) is fundamentally challenged by the extensive range of degradation types and degrees that impact model generalization. Recent advancements in diffusion models have made considerable progress i…

Blind Face RestorationContrastive LearningDenoising

Elucidating and Endowing the Diffusion Training Paradigm for General Image Restoration

2025-06-26 · Xin Lu, Xueyang Fu, Jie Xiao, Zihao Fan 외

While diffusion models demonstrate strong generative capabilities in image restoration (IR) tasks, their complex architectures and iterative processes limit their practical application compared to mainstream reconstructi…

Image Restoration

Elucidating the Exposure Bias in Diffusion Models

2023-08-29 · Mang Ning, Mingxiao Li, Jianlin Su, Albert Ali Salah 외

Diffusion models have demonstrated impressive generative capabilities, but their \textit{exposure bias} problem, described as the input mismatch between training and sampling, lacks in-depth exploration. In this paper, w…

AttributeImage Generation

DeeDSR: Towards Real-World Image Super-Resolution via Degradation-Aware Stable Diffusion

2024-03-31 · Chunyang Bi, Xin Luo, Sheng Shen, Mengxi Zhang 외

Diffusion models, known for their powerful generative capabilities, play a crucial role in addressing real-world super-resolution challenges. However, these models often focus on improving local textures while neglecting…

Contrastive LearningDenoisingImage Super-ResolutionSuper-Resolution