Reproduce “Improving Neural Symbolic Regression with a Diverse Prior and Post-training Alignment” 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
David Otte, Jörg K.H. Franke, Frank Hutter
RASR narrows the gap between amortized symbolic regression and search-based methods by training on realistic synthetic tables and post-training generated expressions for fit and size. On SRBench 2.0’s 12-dataset black-box track, RASR-94M-L has the best mean test-R² rank among evaluated methods, with 7.4% mean improvability. The result is limited to a small benchmark, where four datasets are reduced to meet input limits.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.