Reproduce “Algorithm Modelization: Transforming Optimization Algorithms into Evolvable Intelligence” 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
Jian Yao, Zhiyao Zhang, Yichen Li, Xingyu Wu, Kay Chen Tan
A2M uses search histories to generate candidate populations, learning from existing optimizers before refining its search through performance feedback. It conditions proposals on recent population and fitness states, evaluates candidates, and retains improvements. Across 120 continuous BBOB configurations, it achieved average rank 1.52 versus 4.60 for GTDE. Continuous and mixed-variable settings require separate trained instances.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.