MedSAM and EchoJEPA: Repurposing Meta's General AI Models for Medical Breakthroughs
BoWang87 · x · 2026-10-23
- SAM → MedSAM: adapting Meta's general image segmentation model to identify anatomy and disease across medical imaging modalities, trained on 1.5M+ image-mask pairs.
- JEPA → EchoJEPA: building on video representation learning to study the beating heart, trained on 18M ultrasound videos.
Presenting at Meta's Open Science Summit, Bo Wang argued the next medical AI breakthrough may come from models never built for medicine, crediting Meta's open-source strategy for giving researchers foundations far beyond the models' original purpose.
More from Models
- DeepSeek V4.1 Flash matches Gemini 3.1 Pro on ARC-AGI-2 at ~1/6 the cost — teortaxesTex · 2026-10-23
- Smaug to integrate Nvidia Nemotron: 'run a powerful LLM from one machine' teased — bindureddy · 2026-10-23
- Claim: Opus 5.5 quietly nerfed to make the next model look like a bigger jump — RileyRalmuto · 2026-10-23
- Three small MoEs tested in one agent pipeline: Flint 40B fastest and cheapest, Nemotron chokes on edits — GapNew4766 · 2026-10-23
- Abliterated GLM-5.3 open-weight model released with 2x cyber exploitation capability — AaronBergman18 · 2026-10-23
- Artificial Analysis launches Coding Agent Index v2.0 with lower scores across the board — ArtificialAnlys · 2026-10-23