Penguins Don't Fly: Reasoning about Generics through Instantiations and Exceptions
This addresses the challenge of handling exceptions in generic knowledge for NLP applications, though it is incremental as it builds on existing linguistic theory and benchmarks.
The paper tackled the problem of generating specific cases (exemplars) for generic statements, which are crucial for understanding when such statements hold true or false, by presenting a novel framework informed by linguistic theory that outperformed a GPT-3 baseline by 12.8 precision points.
Generics express generalizations about the world (e.g., birds can fly) that are not universally true (e.g., newborn birds and penguins cannot fly). Commonsense knowledge bases, used extensively in NLP, encode some generic knowledge but rarely enumerate such exceptions and knowing when a generic statement holds or does not hold true is crucial for developing a comprehensive understanding of generics. We present a novel framework informed by linguistic theory to generate exemplars -- specific cases when a generic holds true or false. We generate ~19k exemplars for ~650 generics and show that our framework outperforms a strong GPT-3 baseline by 12.8 precision points. Our analysis highlights the importance of linguistic theory-based controllability for generating exemplars, the insufficiency of knowledge bases as a source of exemplars, and the challenges exemplars pose for the task of natural language inference.