LuMamba fixes cross-hospital EEG AI failure, boosts Alzheimer's detection by 20%
alex_verem · x · 2026-08-14
Researchers from ETH Zurich and University of Bologna developed LuMamba to solve the problem of AI EEG diagnostic tools failing across hospitals. By translating electrode layouts to a standard format, it achieves high accuracy across institutions. On Alzheimer's detection, LuMamba scores 0.97 AUPR on unseen electrode layouts, a 20% improvement over standard training and 4% over previous SOTA. It also uses 377x fewer computations than LaBraM and handles 12x longer recordings.
More from Research
- Netflix details its production LLM judge: hundreds of thousands of recommendations scored weekly — omarsar0 · 2026-08-24
- Nature Comment: Provenance, not interpretability, grounds trust in autonomous science — gabepgomes · 2026-08-24
- New Architecture RHEA: Train 1B Model on 8GB VRAM — zemondza · 2026-08-24
- Trained two 16M-param models to do generative CAD with real physics — debreuil · 2026-08-24
- Claude model helps discover complex structure on S^6, solving 60-year-old math problem — Singularitarian · 2026-08-24
- Study: Agents read instructions/notes 60.5% of the time, rarely touch API docs — dair_ai · 2026-08-24