SURGE Surgical Foundation Model: Perception Saturates After Just One Labeled Video
maier_ak · x · 2026-10-14
The CAMMA group in Strasbourg built SURGE, a surgical foundation model pretrained on 30M+ surgical frames, then systematically measured where more labels stop helping.
Key finding: perception saturates at a single labeled video — segmentation hits 76.68 mIoU and stays flat up to eight labeled videos. This thread covers the perception side; the author's follow-up post adds reasoning-task results.
Related event: Surgical foundation model SURGE maps the limits of annotation gains(3 posts)→
More from Research
- IBM ported Spyre's torch.compile backend to Inductor, cutting code 7x — lmoroney · 2026-10-14
- Neuro for AI & AI for Neuro workshop at AAAI 2027 opens call for papers in Montreal — naturecomputes · 2026-10-14
- BVB Benchmark Tests Agent Video Understanding by Rebuilding 288 Real Videos in Blender — SonglinYang4 · 2026-10-14
- NSF awards $20M to AI Institute in Dynamic Systems for robotic manufacturing — eigensteve · 2026-10-14
- Should ICLR slash its acceptance rate to ~10% as submissions surge? — boknilev · 2026-10-14
- Is 'Attention' the Essence of Intelligence? Quanta examines the idea behind Transformers — TMWNN · 2026-10-14