Canonical Knowledge Selection is a scholarly contribution at the paper level—such as a method, model, framework, theory, construct, dataset, task, or metric—that is directly situated under its parent topic and represents a foundational or standard advancement within that domain.