Incorporating Commonsense Knowledge Graph in Pretrained Models for Social Commonsense Tasks
This addresses the need for better social intelligence in AI systems, particularly for tasks requiring commonsense reasoning, though it is incremental as it builds on existing methods for knowledge infusion.
The paper tackled the problem of improving social commonsense reasoning in pretrained language models by incorporating external commonsense knowledge graphs like ConceptNet, achieving strong performance on the SocialIQA task in both limited and full data settings.
Pretrained language models have excelled at many NLP tasks recently; however, their social intelligence is still unsatisfactory. To enable this, machines need to have a more general understanding of our complicated world and develop the ability to perform commonsense reasoning besides fitting the specific downstream tasks. External commonsense knowledge graphs (KGs), such as ConceptNet, provide rich information about words and their relationships. Thus, towards general commonsense learning, we propose two approaches to \emph{implicitly} and \emph{explicitly} infuse such KGs into pretrained language models. We demonstrate our proposed methods perform well on SocialIQA, a social commonsense reasoning task, in both limited and full training data regimes.