Scholars Question Evaluation Standards in Single-Cell Perturbation Prediction
anshulkundaje · x · 2026-07-08
Stanford scholar Anshul Kundaje points out severe flaws in the experimental design and performance metrics of current single-cell perturbation prediction models. He warns that if a model can easily beat those trained on massive datasets, it indicates fundamental deviations in the field's evaluation benchmarks and timeline settings.
More from Research
- Does LeWorldModel Preserve Local Action Geometry? Gram Matrix Analysis Says Barely — Vasil Vasilev · 2026-10-26
- Busy Beaver explainer video hits 100k views in one day — Paimaamu · 2026-10-26
- Toggling a Single Neuron Turns Qwen into Cookie Monster—Plus a Zoo Map — burny_tech · 2026-10-26
- A neural net learns 2^100 LDPC codewords from just 10-20 examples — DimaKrotov · 2026-10-26
- AGMAI advises AI labs against designating mathematicians to vet model math outputs — AlexKontorovich · 2026-10-26
- 96-Head MLA Hits 363 FLOPs/Byte Compute Intensity; FP8 KV Still Can't Escape Compute Bound — YouJiacheng · 2026-10-26