Relevant CommonSense Subgraphs for "What if..." Procedural Reasoning
This addresses a challenge in AI for natural language processing, but it is incremental as it builds on existing methods for knowledge graph reasoning.
The paper tackles the problem of answering 'What if...' questions in procedural text by using external commonsense knowledge, achieving state-of-the-art performance on the WIQA benchmark.
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 evaluate our model on WIQA benchmark and achieve state-of-the-art performance compared to the recent models.