Reproduce “Magnolia: A Persistent-Memory Agent Framework for Computational Chemistry” 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
Tao Jiang, Benjamin Navet, Yousouf Taghzouti, Fabien Gandon, Louis-Félix Nothias, Élise Dumont
Magnolia carries project knowledge across computational-chemistry sessions by recording runs, turning observations into reviewable notes, and using relevant notes and human-approved rules later. In protein–peptide docking, researchers used this history to identify that better HADDOCK3 scores came from restraints targeting the wrong pocket. In a four-task replay, memory changed answers, though evidence was limited to one project.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.