[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-a-smarter-order-for-ai-image-generation":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},3599,"a-smarter-order-for-ai-image-generation","A Smarter Order for AI Image Generation","Researchers propose a spanning-tree traversal method that keeps autoregressive image models fast while making in-painting and editing genuinely flexible.","A new training approach lets autoregressive image models handle flexible editing without paying the usual speed penalty.\n\nMost autoregressive visual models generate images patch by patch in a fixed left-to-right order — efficient, but rigid. Researchers behind STAR (Spanning Tree Autoregressive) modeling instead route generation through traversal orders derived from uniform spanning trees built on a lattice of image patch positions. Breadth-first search over those trees ensures that whatever portion of an image has already been generated appears as a contiguous prefix, which makes native in-painting straightforward rather than bolted on. The method uses rejection sampling to pick trees whose traversal properties satisfy that constraint, requiring no meaningful changes to standard language-style autoregressive architectures.\n\nThe significance is in the tradeoff it sidesteps. Prior attempts to give autoregressive image models arbitrary sequence flexibility — typically via random permutation during training — tend to degrade sample quality, hurt inference speed, or both. STAR threads that needle by using structured randomization rather than pure randomness, preserving what researchers call \"postfix completion\" capability while baking in image priors like center bias and locality. That means one model can plausibly handle generation and editing tasks that currently require separate specialized systems.\n\nAutoregressive visual generation has been chasing diffusion models on quality benchmarks for two years; the gap has narrowed but the flexibility argument usually still favors diffusion. STAR does not claim to close that gap outright, but if structured sequence orders genuinely hold up at scale, the architectural simplicity — no diffusion process, no separate in-painting model — becomes a practical argument, not just an academic one.","[\"ai\",\"image-generation\",\"research\",\"autoregressive-models\"]","2026-07-03T04:00:00.000Z","2026-07-03T09:19:35.568Z","2026-07-03T09:19:38.527Z","published",null,[],"ai",[24,26,27,28],"image-generation","research","autoregressive-models",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2511.17089",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"]