Prompt tuning involves optimizing the learnable prompt to improve the computational efficiency and accuracy of downstream tasks. It is based on the similarity of subgraphs and is formulated using prompt-assisted task-specific subgraph representations.

→ View original post on X — @maximelabonne