Reproduce “Evolutionary Feature Engineering for Structured Data” 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
Ege Onur Taga, Yilin Zhuang, M. Emrullah Ildiz, Petros Mol, Abhimanyu Das, Karthik Duraisamy, Samet Oymak
The authors present Evolutionary Feature Engineering, where an LLM evolves preprocessing programs scored by a fixed downstream model. For time series, programs transform history and invert forecasts; for tabular data, they engineer features. EFE-Time reduces mean MASE by 3.0% on 10 selected forecasting datasets; EFE-Tab ranks 1.39 among four methods on nine selected classification datasets with a single tree. Results are dataset-limited.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.