Incorporating External Knowledge to Enhance Tabular Reasoning
This addresses the problem of enhancing tabular reasoning for NLP applications, but it appears incremental as it focuses on modifications to existing approaches.
The paper tackled the challenge of tabular natural language inference by proposing easy modifications to how information is presented to models, resulting in substantial performance improvements as shown through systematic experiments.
Reasoning about tabular information presents unique challenges to modern NLP approaches which largely rely on pre-trained contextualized embeddings of text. In this paper, we study these challenges through the problem of tabular natural language inference. We propose easy and effective modifications to how information is presented to a model for this task. We show via systematic experiments that these strategies substantially improve tabular inference performance.