A review finds many medical-imaging papers still split data in leak-prone ways
wandedob · x · 2026-07-27
The post explains that the report checks whether medical-imaging papers split data correctly at the patient or whole-volume level, or incorrectly at the slice/patch level, which causes leakage and inflated metrics.
It covers papers using 2D CNNs or ViTs on 3D medical volumes such as MRI, CT, OCT, and ultrasound, as well as pathology whole-slide images broken into patches.
Related event: AI Agent Audit Reveals Data Leakage in Medical Imaging Papers(4 posts)→
More from Research
- Mathematician shares a cheap 4-step heuristic for hyperparameter tuning — dejanseo · 2026-09-23
- Burkov skew AI hype: 'deterministic LLMs' and 'first agents' are old tricks rebranded — burkov · 2026-09-23
- Continuous diffusion beats discrete on random k-SAT, proposed as standard benchmark — ArashVahdat · 2026-09-23
- Grady Booch: Contemporary AI Still Lacks Abductive Reasoning, Just 'Next-Token Prediction' — Grady_Booch · 2026-09-23
- AI solves Navier-Stokes-related problem as machines upend mathematics, New Scientist reports — burny_tech · 2026-09-23
- Mathematician says OpenAI likely proved a significant partial case of the Hodge conjecture — burny_tech · 2026-09-23