Possible Stories: Evaluating Situated Commonsense Reasoning under Multiple Possible Scenarios
Possible Stories: Evaluating Situated Commonsense Reasoning under Multiple Possible Scenarios
复制标题
DOI：
10.48550/arxiv.2209.07760
复制
发表时间：
2022-09
期刊：
影响因子：
--
通讯作者：
Mana Ashida;Saku Sugawara
中科院分区：
文献类型：
--
作者：
Mana Ashida;Saku Sugawara
相同上下文的可能后果可能会根据我们所指的情况而有所不同。可能的结局是候选人的答案，鉴于我们由此产生的数据集。 ，强调在无监督的环境中的最高准确性（60.2％）远远落后于人类的准确性（92.5％） ，我们的数据集包括需要反事实推理的示例，以及需要读者的反应和虚构信息的示例，这表明我们的数据集可以作为对未来常识性推理的未来研究的挑战。
The possible consequences for the same context may vary depending on the situation we refer to. However, current studies in natural language processing do not focus on situated commonsense reasoning under multiple possible scenarios. This study frames this task by asking multiple questions with the same set of possible endings as candidate answers, given a short story text. Our resulting dataset, Possible Stories, consists of more than 4.5K questions over 1.3K story texts in English. We discover that even current strong pretrained language models struggle to answer the questions consistently, highlighting that the highest accuracy in an unsupervised setting (60.2%) is far behind human accuracy (92.5%). Through a comparison with existing datasets, we observe that the questions in our dataset contain minimal annotation artifacts in the answer options. In addition, our dataset includes examples that require counterfactual reasoning, as well as those requiring readers’ reactions and fictional information, suggesting that our dataset can serve as a challenging testbed for future studies on situated commonsense reasoning.