Relevant CommonSense Subgraphs for "What if..." Procedural Reasoning [article]

Chen Zheng, Parisa Kordjamshidi
2022 arXiv   pre-print
We study the challenge of learning causal reasoning over procedural text to answer "What if..." questions when external commonsense knowledge is required. We propose a novel multi-hop graph reasoning model to 1) efficiently extract a commonsense subgraph with the most relevant information from a large knowledge graph; 2) predict the causal answer by reasoning over the representations obtained from the commonsense subgraph and the contextual interactions between the questions and context. We
more » ... uate our model on WIQA benchmark and achieve state-of-the-art performance compared to the recent models.
arXiv:2203.11187v2 fatcat:ol4kqpzjhzetxi2f3npgbfirt4