🎉 Our paper “Self-supervised Blending for Molecular Sequence Generation” was accepted by NeurIPS 2025! We introduce a blending-based SSL strategy to surmount the data scarcity challenge in molecular design.