GeneICL: lightweight tabular foundation model wins across 80 clinical transcriptomics tasks
anshulkundaje · x · 2026-10-07
A Stanford team (Anshul Kundaje's group) introduces GeneICL, a lightweight looped tabular foundation model for bulk transcriptomics, benchmarked on 80 classification, regression, and survival-prediction tasks. Key insight: self-supervised transcriptomic FMs pretrained on millions of profiles often lose to simple supervised baselines — the pretraining objective may matter more than scale.
Related event: GeneICL: Lightweight Transcriptomics Model Beats Larger Baselines(2 posts)→
More from Research
- RWTH Aachen's ARROW unifies 3D reconstruction and point tracking from any RGB inputs, sets new SOTA — kwangmoo_yi · 2026-10-09
- HAIPS 2026 workshop lands at COLM tomorrow with top AI privacy researchers — tianshi_li · 2026-10-09
- Tighter Bayes error bounds via generalized Bhattacharyya and Chernoff means — FrnkNlsn · 2026-10-09
- PAMI: part-anchored motion lifts contact recall 14.5% in text-to-HOI generation — _akhaliq · 2026-10-09
- Study finds LLM novelty judges are unstable: scores swing wildly with evaluation design choices — _akhaliq · 2026-10-09
- Google's AMIE lands in The Lancet: 90% match with doctors' final diagnoses in real clinics — sundarpichai · 2026-10-09