[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-benchmark-exposes-limits-of-ai-models-simulating-aneurysms":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},9256,"benchmark-exposes-limits-of-ai-models-simulating-aneurysms","Benchmark Exposes Limits of AI Models Simulating Aneurysms","A new dataset of aneurysm blood-flow simulations shows AI models improve with synthetic geometry variants, but accuracy frays over longer simulations.","AI models that predict blood flow in brain aneurysms get a boost from training on computer-generated shapes, but that boost doesn't survive longer simulations.\n\nA team of researchers built AneumoBench, a dataset linking 401 real aneurysm geometries to 9,693 \"descendant\" variants generated through local shape editing, with fluid-dynamics simulations run on all of them. The dataset includes 80,752 steady-state velocity and pressure cases across eight different inlet conditions, plus 9,715 time-based sequences that track velocity, pressure, and wall shear stress across a simulated one-second heartbeat, sampled in 100 frames. The researchers trained nine neural network architectures three ways: on real geometries only, on synthetic descendants only, and on descendants first followed by fine-tuning on real ones. Across 79 held-out real geometries never seen in training, the two-stage approach beat training on real data alone for steady flow fields and for short-term wall shear stress predictions.\n\nThe catch shows up when models have to look further ahead. Given just four observed frames, models predict wall shear stress forward across the remaining 96 steps of the sequence, and the accuracy gains from synthetic pretraining fade well before that rollout finishes. Lower trajectory error also didn't reliably translate into correctly flagging where stress concentrates, the detail that actually predicts rupture risk.\n\nThat's a useful reality check for a field leaning on synthetic data because real aneurysm scans are scarce: a model that nails the first few frames isn't automatically one a clinician should trust for the whole cycle.","[\"ai\",\"machine-learning\",\"healthcare\",\"benchmarks\"]","2026-10-01T04:00:00.000Z","2026-10-02T05:08:39.560Z","2026-10-02T05:08:43.586Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"publisher-r1","publisher",1,"The body states transient simulations track 100 frames of the cardiac cycle but later refers to a 'full 96-step cardiac cycle,' an internal numeric inconsistency.","resolved","ai",[30,32,33,34],"machine-learning","healthcare","benchmarks",[36],{"name":37,"url":38},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.14717",0,{"sections":41},[42,45,49,53,58,63,67,72,77,81,86,91,96,101],{"name":43,"slug":30,"count":44,"latest_published_at":18},"AI",5659,{"name":46,"slug":47,"count":48,"latest_published_at":18},"Security","security",818,{"name":50,"slug":51,"count":52,"latest_published_at":18},"Policy","policy",430,{"name":54,"slug":55,"count":56,"latest_published_at":57},"Deals","deals",298,"2026-09-30T21:00:26.000Z",{"name":59,"slug":60,"count":61,"latest_published_at":62},"Hardware","hardware",196,"2026-09-30T13:00:00.000Z",{"name":64,"slug":65,"count":66,"latest_published_at":18},"Science","science",163,{"name":68,"slug":69,"count":70,"latest_published_at":71},"Consumer Tech","consumer-tech",149,"2026-09-30T22:57:11.000Z",{"name":73,"slug":74,"count":75,"latest_published_at":76},"Dev Tools","dev-tools",93,"2026-10-01T02:30:48.000Z",{"name":78,"slug":79,"count":75,"latest_published_at":80},"Software","software","2026-09-30T21:41:11.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Startups","startups",84,"2026-09-30T20:39:09.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"Gaming","gaming",51,"2026-09-30T16:24:30.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":102,"slug":103,"count":104,"latest_published_at":105},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]