Stanford's Kundaje challenges cell-line prediction method over train-test similarity concerns
anshulkundaje · x · 2026-11-20
Anshul Kundaje questioned whether a cell-line drug-response prediction method is truly invariant to train-test similarity, suspecting a weighted-means trick or structural exploit. The authors rebutted that they held out entire classes of cell lines, so no matches existed in the test set.
More from Research
- MIT's Hiroshi Ishii shares Tangible Media Group's Creative Tension principles — ishii_mit · 2026-11-20
- Train a real LLM entirely in your browser with this interactive demo — morgymcg · 2026-11-20
- Ex-OpenAI Researcher: English Is the New ML Programming Language, Lean the Type System — lukaszkaiser · 2026-11-20
- S1CogBench probes System One decision models across 24 cognitive tasks — Dengzhe Hou · 2026-11-20
- AFFECT from KAIST makes world models dynamics-faithful via action effect matching — John Won · 2026-11-20
- Mind2Act World: a 180-task benchmark exposing robots' mind-action composition failures — Mind2Act-World · 2026-11-20