论文标题
使用矛盾改善了问答系统
Using contradictions improves question answering systems
论文作者
论文摘要
这项工作研究了自然语言推论（NLI）中矛盾的使用来回答（QA）。通常，NLI系统通过确定潜在的答案是否为\ emph {intailed}（由某些背景上下文支持）来帮助回答问题。但是，还可以确定答案是否与上下文相矛盾吗？我们在两个设置（多项选择和提取质量检查）中对此进行了测试，并发现合并矛盾的系统可以比某些数据集中的仅限额系统更好地做得更好。但是，最佳性能来自使用矛盾，元素和质量检查模型置信度得分。这对在安全性是一个问题的范围内的质量检查系统的部署有影响。
This work examines the use of contradiction in natural language inference (NLI) for question answering (QA). Typically, NLI systems help answer questions by determining if a potential answer is \emph{entailed} (supported) by some background context. But is it useful to also determine if an answer contradicts the context? We test this in two settings, multiple choice and extractive QA, and find that systems that incorporate contradiction can do slightly better than entailment-only systems on certain datasets. However, the best performances come from using contradiction, entailment, and QA model confidence scores together. This has implications for the deployment of QA systems in domains such as medicine and science where safety is an issue.
