4 papers - avg viability 6.3
FlowMS uses discrete flow matching to generate novel molecular structures from mass spectrometry data, outperforming existing methods and enabling faster discovery in metabolomics and natural product research.
M^4olGen offers a novel framework for generating molecules under precise property constraints, outperforming existing models.
Develop a cutting-edge molecular graph generation tool using hierarchical discrete diffusion models for enhanced drug discovery and material science applications.
A new benchmark and dataset for training reasoning LLMs to generate novel, high-quality, and diverse molecules for drug discovery.