{"slug": "resolving-endpoint-underfitting-in-diffusion-bridges-via-noise-alignment", "title": "Resolving Endpoint Underfitting in Diffusion Bridges via Noise Alignment", "summary": "Researchers have identified a critical underfitting problem in diffusion bridge models near the target endpoint, caused by a mismatch in noise levels between network input and regression target. The team proposes the Noise-Aligned Diffusion Bridge (NADB), which reformulates the diffusion process using a mean network and noise-aligned mapping to correct this anomaly. Experimental results across image restoration and translation tasks confirm the method's effectiveness in resolving endpoint underfitting.", "body_md": "arXiv:2605.28962v1 Announce Type: new\nAbstract: 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.", "url": "https://wpnews.pro/news/resolving-endpoint-underfitting-in-diffusion-bridges-via-noise-alignment", "canonical_source": "https://arxiv.org/abs/2605.28962", "published_at": "2026-05-29 04:00:00+00:00", "updated_at": "2026-05-29 04:15:28.180220+00:00", "lang": "en", "topics": ["machine-learning", "generative-ai", "computer-vision", "neural-networks", "ai-research"], "entities": ["Noise-Aligned Diffusion Bridge", "NADB"], "alternates": {"html": "https://wpnews.pro/news/resolving-endpoint-underfitting-in-diffusion-bridges-via-noise-alignment", "markdown": "https://wpnews.pro/news/resolving-endpoint-underfitting-in-diffusion-bridges-via-noise-alignment.md", "text": "https://wpnews.pro/news/resolving-endpoint-underfitting-in-diffusion-bridges-via-noise-alignment.txt", "jsonld": "https://wpnews.pro/news/resolving-endpoint-underfitting-in-diffusion-bridges-via-noise-alignment.jsonld"}}