I Know What You Asked: Graph Path Learning using AMR for Commonsense Reasoning

• DOI: 10.18653/v1/2020.coling-main.222
• Publication Date: 2021-01-08T13:58:31Z
• Authors: Jungwoo Lim; Dongsuk Oh; Yoonna Jang; Kisu Yang; Heuiseok Lim
• Keywords: Commonsense reasoning; Commonsense knowledge

CommonsenseQA is a task in which correct answer predicted through commonsense reasoning with pre-defined knowledge. Most previous works have aimed to improve the performance distributed representation without considering process of predicting from semantic question. To shed light upon interpretation question, we propose an AMR-ConceptNet-Pruned (ACP) graph. The ACP graph pruned full integrated encompassing Abstract Meaning Representation (AMR) generated input questions and external knowledge graph, ConceptNet (CN). Then exploited interpret path as well predict on task. This paper presents manner can be interpreted relations concepts provided by Moreover, ACP-based models are shown outperform baselines.