Brain-IT from Weizmann reconstructs what you're seeing by reading brain activity, with colorful mistakes
Brighter-Side-News · reddit · 2026-10-09
Michal Irani's lab at the Weizmann Institute built Brain-IT, an AI system that reconstructs surprisingly detailed images from the brain activity of someone viewing them — though it errs, turning a dog in a bathtub into a similarly colored goat and a cake into three sandwiches. The study appeared at ICLR 2026, with related universal brain encoder work at CCN.
The two projects tackle both directions: predicting brain responses to images and reconstructing images from those responses. Demonstrations cover only pictures viewed during scanning; dream reading and patient communication remain research ambitions.
More from Research
- Embedding arithmetic demo: cartoon − animation − comic = diagram — mayfer · 2026-10-23
- Domingos Cites New Result Claiming RL Is 'Basically Useless' vs Supervised Learning — pmddomingos · 2026-10-23
- Wolfram Opens Rule 30 Prizes to AIs: $30,000 for Three Open Problems — cephaloform · 2026-10-23
- Stanford and Databricks use AI agents to reproduce medical findings and test analytical robustness — ChengleiSi · 2026-10-23
- EMNLP paper: diffusion LMs internally encode latent denoising timestep representations — gsarti_ · 2026-10-23
- KU Leuven and VIB open two Group Leader professorships in AI for Biology, start 2027 — jpillowtime · 2026-10-23