"Law and literature " comes from an American radical law school of the1970 s,and it was then considered a campaign,thus the school became one of the most post-modernistic schools of thought. Due to its weste..."Law and literature " comes from an American radical law school of the1970 s,and it was then considered a campaign,thus the school became one of the most post-modernistic schools of thought. Due to its western theoretic background,it is necessary to interpret its context; in other words,the "Law and literature"problems shall be interpreted in the Chinese context,which can also provide much inspiration and reference for the construction of China's legal system.展开更多
随着人工智能技术的发展和海量司法数据的公开,面向“智慧司法”服务的司法判决预测(legal judgment prediction,LJP)任务受到了学术界和工业界的广泛关注,该任务旨在根据有限的案件事实描述文本来预测案件的罪名、法条和刑期。然而,现...随着人工智能技术的发展和海量司法数据的公开,面向“智慧司法”服务的司法判决预测(legal judgment prediction,LJP)任务受到了学术界和工业界的广泛关注,该任务旨在根据有限的案件事实描述文本来预测案件的罪名、法条和刑期。然而,现有工作缺乏对易混淆司法案件的智能决策的研究,且相关模型通常缺乏可解释性,这会导致模型预测严重依赖领域专家,阻碍LJP在不同法律体系中的应用。为此,提出了一种基于因果图分析的司法判决预测(prediction of legal judgment based on causal graph analysis,CGLJ)方法,首先从非结构化的法律事实描述文本中挖掘要素之间的因果关系,然后采用易混淆罪名聚类的构图方法构建因果图,既考虑了相似事实描述之间的差异,又增强了事实描述和法律法规之间的相互作用,最后将构建好的因果图融入深度神经网络进行联合推理,得到判决预测结果。此外,还对模型预测过程中的因果图推理过程进行了可视化,为判决结果提供了更好的可解释性。在2018中国“法研杯”司法人工智能挑战赛(CAIL2018)司法判决预测数据集上的实验结果表明,该方法相比基线模型取得了更好的效果。展开更多
文摘"Law and literature " comes from an American radical law school of the1970 s,and it was then considered a campaign,thus the school became one of the most post-modernistic schools of thought. Due to its western theoretic background,it is necessary to interpret its context; in other words,the "Law and literature"problems shall be interpreted in the Chinese context,which can also provide much inspiration and reference for the construction of China's legal system.
文摘随着人工智能技术的发展和海量司法数据的公开,面向“智慧司法”服务的司法判决预测(legal judgment prediction,LJP)任务受到了学术界和工业界的广泛关注,该任务旨在根据有限的案件事实描述文本来预测案件的罪名、法条和刑期。然而,现有工作缺乏对易混淆司法案件的智能决策的研究,且相关模型通常缺乏可解释性,这会导致模型预测严重依赖领域专家,阻碍LJP在不同法律体系中的应用。为此,提出了一种基于因果图分析的司法判决预测(prediction of legal judgment based on causal graph analysis,CGLJ)方法,首先从非结构化的法律事实描述文本中挖掘要素之间的因果关系,然后采用易混淆罪名聚类的构图方法构建因果图,既考虑了相似事实描述之间的差异,又增强了事实描述和法律法规之间的相互作用,最后将构建好的因果图融入深度神经网络进行联合推理,得到判决预测结果。此外,还对模型预测过程中的因果图推理过程进行了可视化,为判决结果提供了更好的可解释性。在2018中国“法研杯”司法人工智能挑战赛(CAIL2018)司法判决预测数据集上的实验结果表明,该方法相比基线模型取得了更好的效果。