[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-method-trims-ai-image-encoder-computation-by-72-percent":10,"sections":40},{"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":30,"tags":31,"sources":35,"feedback":39,"feedback_at":22,"cost_usd":39,"total_tokens":39},8296,"new-method-trims-ai-image-encoder-computation-by-72-percent","New Method Trims AI Image Encoder Computation by 72 Percent","A new selective-evaluation technique cuts encoder computation by about 72% while improving image reconstruction quality at low bit rates.","AI researchers have found a way to make image-compression systems smarter about where they spend computing power, without sacrificing image quality.\n\nThe system, called ACV-Gate, tackles a bottleneck in generative image communication: sending compact \"tokens\" that a receiver reconstructs into a full image under a tight data budget. Picking the best tokens usually means the sender has to simulate the receiver's reconstruction for every candidate, which is computationally expensive. ACV-Gate instead trains a model to predict which candidates matter most, then spends full evaluation effort only on that shortlist. In tests on CIFAR-10, the technique improved image quality (PSNR) by 0.636 dB over a baseline method called LocalMDL at a low bit rate of 0.20 bits per pixel, while running only 2.13 full evaluations per image on average, 27.60% of the calls needed by an exhaustive baseline method - a 72% cut in encoder-side computation. The researchers also tested it on STL-10 and larger 384x384 images with similar results.\n\nThat computation savings matters most where bandwidth and processing power are both scarce, like video calls over weak connections, satellite links, or drones. It also echoes a pattern showing up across AI: rather than brute-forcing every option, systems increasingly learn to guess which paths are worth fully checking, similar to how speculative decoding speeds up large language models. If it holds up outside the lab, this kind of selective computation could matter more for real-time image and video transmission than another modest bump in raw compression ratio.\n\nStill, this is a preprint tested on small academic datasets like CIFAR-10, not a shipped product. Getting from a clever training trick to actual bandwidth savings on your video call is a longer road than the abstract suggests.","[\"ai\",\"image-compression\",\"neural-networks\",\"research\"]","2026-09-28T04:00:00.000Z","2026-09-28T22:23:11.996Z","2026-09-28T22:23:19.426Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"Fix the dek's claim of 'cutting encoder computation by nearly 75%' to match the body's own figure of a 72% cut (27.60% of Exact-Full calls means a 72.4% reduction, not nearly 75%), since the dek's framing overstates what the body itself states.","resolved","ai",[30,32,33,34],"image-compression","neural-networks","research",[36],{"name":37,"url":38},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.30756",0,{"sections":41},[42,46,51,56,61,66,71,76,81,86,91,96,101,106],{"name":43,"slug":30,"count":44,"latest_published_at":45},"AI",4900,"2026-09-28T17:44:43.000Z",{"name":47,"slug":48,"count":49,"latest_published_at":50},"Security","security",766,"2026-09-28T15:35:23.000Z",{"name":52,"slug":53,"count":54,"latest_published_at":55},"Policy","policy",405,"2026-09-28T17:00:51.000Z",{"name":57,"slug":58,"count":59,"latest_published_at":60},"Deals","deals",269,"2026-09-28T17:42:46.000Z",{"name":62,"slug":63,"count":64,"latest_published_at":65},"Hardware","hardware",191,"2026-09-28T15:45:00.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Science","science",154,"2026-09-28T13:19:18.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Consumer Tech","consumer-tech",139,"2026-09-28T17:09:47.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Software","software",91,"2026-09-25T20:55:00.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Dev Tools","dev-tools",87,"2026-09-28T16:11:42.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"Startups","startups",80,"2026-09-28T17:50:28.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"General","general",49,"2026-09-28T16:44:57.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"Gaming","gaming",48,"2026-09-25T18:35:21.000Z",{"name":102,"slug":103,"count":104,"latest_published_at":105},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":107,"slug":108,"count":109,"latest_published_at":110},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]