Reproduce “Model-Agnostic Feature Selection via LOCO-Guided Adaptive Minipatch Sampling” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xuhui Liu, Lili Zheng
LAMPS selects predictive features by repeatedly training regression learners on minipatches, estimating leave-one-covariate-out importance, and adapting feature-sampling probabilities while retaining exploration. In a correlated linear simulation, these probabilities separated signal and noise features more clearly than raw importance scores. In Riboflavin data, LAMPS with Lasso selected 3 features versus 41 and 73 for competing methods, without establishing causal truth.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.