Reproduce “Pre-training of Molecular GNNs via Conditional Boltzmann Generator” 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
Daiki Koge, Naoaki Ono, Shigehiko Kanaya
Boltzmann GNN pretrains a graph neural network to transfer information about molecular conformational ensembles to property prediction from 2D graphs. It conditions a diffusion model on a representation and learns from conformations, incorporating Boltzmann-distribution information during pretraining. In five regression tasks, it achieved the best performance on four, a small-task transfer result, not evidence of runtime savings or distribution fidelity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.