Reproduce “Evolutionary One-Step Generators: Fast and Diverse Sampling for Discrete Design” 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
Marcus Vukojevic, Erik Nielsen, Veronica Lachi, Andrea Passerini, Giovanni Iacca
EGO trains a one-step discrete generator by scoring completed hard outputs and using evolutionary parameter search; at deployment, it proposes each candidate in one network evaluation. On 100 scaffold queries, it produced a mean of 547.3 filter-passing proposals per query versus 54.5 for MoLeR at matched deadlines, but these are computationally filtered proposals, not validated drug candidates.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.