compositional generalization

Compositional generalization is the ability of a model to understand and generate new combinations of known components or concepts. It's crucial for achieving flexible and robust reasoning in AI systems, especially in natural language processing and computer vision.

11 papers