{"id":9694,"date":"2026-01-13T07:02:43","date_gmt":"2026-01-13T07:02:43","guid":{"rendered":"https:\/\/mailitics.com\/index.php\/2026\/01\/13\/2601-06514\/"},"modified":"2026-01-13T07:02:43","modified_gmt":"2026-01-13T07:02:43","slug":"2601-06514","status":"publish","type":"post","link":"https:\/\/mailitics.com\/index.php\/2026\/01\/13\/2601-06514\/","title":{"rendered":"Inference-Time Alignment for Diffusion Models via Doob&#8217;s Matching"},"content":{"rendered":"\n<div>Inference-Time Alignment for Diffusion Models via Doob&#8217;s Matching<\/div>\n<p> \t<BR><br \/>\n<BR><\/BR><br \/>\n    <!-- no image --><br \/>\n \t<BR><br \/>\n<BR><\/BR><\/p>\n<div>arXiv:2601.06514v1 Announce Type: new<br \/>\nAbstract: Inference-time alignment for diffusion models aims to adapt a pre-trained diffusion model toward a target distribution without retraining the base score network, thereby preserving the generative capacity of the base model while enforcing desired properties at the inference time. A central mechanism for achieving such alignment is guidance, which modifies the sampling dynamics through an additional drift term. In this work, we introduce Doob&#8217;s matching, a novel framework for guidance estimation grounded in Doob&#8217;s $h$-transform. Our approach formulates guidance as the gradient of logarithm of an underlying Doob&#8217;s $h$-function and employs gradient-penalized regression to simultaneously estimate both the $h$-function and its gradient, resulting in a consistent estimator of the guidance. Theoretically, we establish non-asymptotic convergence rates for the estimated guidance. Moreover, we analyze the resulting controllable diffusion processes and prove non-asymptotic convergence guarantees for the generated distributions in the 2-Wasserstein distance.<\/div>\n<p> \t<BR><br \/>\n <BR><\/BR><br \/>\n    Jinyuan Chang, Chenguang Duan, Yuling Jiao, Yi Xu, Jerry Zhijian Yang<br \/>\n \t<BR><br \/>\n<BR><\/BR><br \/>\n<a href=\"https:\/\/arxiv.org\/abs\/2601.06514\">Go to original source<\/a><br \/>\n \t<BR><br \/>\n <BR><\/BR><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Inference-Time Alignment for Diffusion Models via Doob&#8217;s Matching arXiv:2601.06514v1 Announce Type: new Abstract: Inference-time alignment for diffusion models aims to adapt a pre-trained diffusion model toward a target distribution without retraining the base score network, thereby preserving the generative capacity of the base model while enforcing desired properties at the inference time. A central mechanism [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[62,113,450,451,376,190,112,191],"tags":[454,4592,900],"class_list":["post-9694","post","type-post","status-publish","format-standard","hentry","category-aimldsaimlds","category-cs-lg","category-cs-na","category-math-na","category-math-oc","category-math-st","category-stat-ml","category-stat-th","tag-diffusion","tag-doob","tag-guidance"],"_links":{"self":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts\/9694"}],"collection":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/comments?post=9694"}],"version-history":[{"count":0,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts\/9694\/revisions"}],"wp:attachment":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/media?parent=9694"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/categories?post=9694"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/tags?post=9694"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}