Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
WeightLoop: Weight-Space Programming with Experimental Feedback
Mengran Li, Bo Li, Weijing Zhao +6
Weight-space programming (WSP) shows that language models can reason about and rewrite neural network weights; recursive self-improvement (RSI) motivates feedback-driven improvemen…
cs.LG2026
CellScientist: From Execution Feedback to Auditable Model-Revision Trajectories for Cellular Perturbation Prediction
Mengran Li, Bo Li, Jiaying Wang +13
Cellular perturbation-response modeling requires coordinated choices of representations, fusion mechanisms, objectives, and training procedures. Large language models (LLMs) can pr…