Candidate Molecule

Candidate molecule research focuses on computationally designing molecules with desired properties for applications like drug discovery and carbon capture. Current efforts leverage machine learning, employing generative models (e.g., GFlowNets, graph transformers, normalizing flows) and hybrid quantum-classical approaches to generate diverse and synthetically accessible molecules, often guided by structure-property relationships learned from large datasets. These advancements promise to accelerate materials discovery and drug development by significantly reducing the time and cost associated with traditional experimental methods.