Activity prediction for low-data scenarios
Generate images from text prompts
Retrieve molecules or images from each other using SMILES or microscopy images