通过噪声对齐解决扩散桥中的端点欠拟合问题
Resolving Endpoint Underfitting in Diffusion Bridges via Noise Alignment
扩散桥模型中目标端点附近的欠拟合问题被揭示，并通过噪声对齐映射解决，可直接用于改进图像复原和翻译任务中的扩散桥训练。
原文依据 we find that this way leads to an anomalous underfitting phenomenon near the target endpoint
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论扩散桥模型为连接两个数据分布（例如在图像复原和图像翻译中）提供了强大的框架。许多现有方法通过模仿标准扩散模型的得分匹配公式来学习这种桥。在这项工作中，我们发现这种方式会导致在目标端点附近（当过程接近目标分布时，t → 0）出现异常欠拟合现象。这种欠拟合表现为预测方差和方向的显著漂移，其原因是网络输入与其回归目标之间的噪声水平差异过大。为了解决这个问题，我们提出了噪声对齐扩散桥（NADB）。我们的方法通过首先使用均值网络提供更干净的条件目标，然后引入一种新颖的噪声对齐映射关系来重新表述扩散桥。这种新表述解决了噪声不匹配问题，并纠正了目标端点附近的欠拟合。在多个图像复原和图像翻译任务上的实验验证了我们方法的有效性。代码可在 https://github.com/gyr02/NADB 获取。
英文原文摘要 1185 字符
Diffusion bridge models offer a powerful framework for connecting two data distributions, such as in image restoration and translation. Many existing methods learn this bridge by mimicking the score-matching formulation of standard diffusion models. In this work, we find that this way leads to an anomalous underfitting phenomenon near the target endpoint, as the process approaches the target distribution (t \to 0). This underfitting, characterized by significant drift in the predicted variance and direction, results from an excessively large discrepancy in noise levels between the network's input and its regression target. To resolve this issue, we propose the Noise-Aligned Diffusion Bridge (NADB). Our approach reformulates the diffusion bridge by first employing a mean network to provide a cleaner conditional target, and then introducing a novel, noise-aligned mapping relationship. This new formulation resolves the noise mismatch and corrects the underfitting near the target endpoint.Experimental validation across multiple image restoration and image translation tasks demonstrates the effectiveness of our approach. Code is available at https://github.com/gyr02/NADB.