Analyzing Commonsense Emergence in Few-shot Knowledge Models

Jeff Da, Ronan Le Bras, Ximing Lu, Yejin Choi, Antoine Bosselut

Keywords: commonsense, knowledge, graph, language, models

TL;DR: We train commonsense knowledge models in few-shot settings to study the emergence of their commonsense representation abilities.

Abstract:

Recently, commonsense knowledge models --- pretrained language models (LM) fine-tuned on knowledge graph (KG) tuples --- showed that considerable amounts of commonsense knowledge can be encoded in the parameters of large language models.
However, as parallel studies show that LMs are poor hypothesizers of declarative commonsense relationships on their own, it remains unclear whether this knowledge is learned during pretraining or from fine-tuning on KG examples.

To investigate this question, we train commonsense knowledge models in few-shot settings to study the emergence of their commonsense representation abilities. Our results show that commonsense knowledge models can rapidly adapt from limited examples, indicating that KG fine-tuning serves to learn an interface to encoded knowledge learned during pretraining. Importantly, our analysis of absolute, angular, and distributional parameter changes during few-shot fine-tuning provides novel insights into how this interface is learned.

Subject Areas: Knowledge Representation, Semantic Web and Search, Question Answering and Reasoning

Archival Status: Archival