[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-daps-speeds-up-diffusion-based-image-restoration":10,"sections":34},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},2814,"daps-speeds-up-diffusion-based-image-restoration","DAPS++ Speeds Up Diffusion-Based Image Restoration","A new solver separates diffusion initialization from reconstruction, cutting compute steps while improving image restoration results.","A research team has rethought how diffusion models solve image reconstruction problems — and the fix is mostly about stopping two tasks from getting in each other's way.\n\nDiffusion-based solvers typically treat image restoration as a joint inference problem, blending a learned prior with measurement data in a single unified process. DAPS++ challenges that framing. The researchers found that the diffusion prior in these solvers isn't doing what the theory says it should — it's mainly acting as a warm-start mechanism that nudges estimates toward the data manifold, while actual reconstruction is driven almost entirely by measurement consistency. Separating those two roles explicitly, rather than letting them interfere, turns out to produce better results with fewer compute steps.\n\nThe efficiency gains matter because diffusion-based inverse solvers are expensive. Every function evaluation costs time and money, and these models are increasingly used in real image restoration pipelines — medical imaging, photography, and satellite imagery among them. A method that gets comparable or better reconstruction quality while requiring fewer steps is directly useful, not just academically interesting.\n\nThe paper is a fourth-revision preprint, which means the core ideas have survived several rounds of scrutiny — though peer review is still the bar that counts.","[\"ai\",\"diffusion-models\",\"image-restoration\",\"research\"]","2026-06-30T04:00:00.000Z","2026-06-30T13:22:40.426Z","2026-06-30T13:22:43.245Z","published",null,[],"ai",[24,26,27,28],"diffusion-models","image-restoration","research",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.17038",0,{"sections":35},[36,40,45,50,55,60,65,70,75,80,85,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":86,"slug":87,"count":83,"latest_published_at":88},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]