人工智能民事司法应用的法律知识图谱构建——以要件事实型民事裁判论为基础

The Construction of Map of Legal Knowledge in Application of Artificial Intelligence in Civil Justice:On the Basis of Essential Facts Theory

在人工智能推动下的司法改革当中,让机器通过深度学习以认知个案,是人工智能司法应用的前提与薄弱之处。当前,地方法院的实践尚未形成有效的知识积累方法。人工智能司法应用的前提是法律知识图谱的构建以及裁判规则的类型化与要素化。要件事实型民事裁判论与司法人工智能的生成规律具有内生契合性,可作为神经网络深度学习、分词设置、知识图谱设计的基础与前端理论。具体应用路径是依要件事实论不断进行层级解构,将案件认事用法解构为不同层级要素,并由法律专家分层级、分阶段标注,从而形成以要素标注的法律知识图谱大数据,以供机器学习。要件事实论之于人工智能司法应用具有独立性,人工智能难以代替法官。人工智能司法应用与民事诉讼制度具有相互促进关系,两者的深度融合将开拓中国民事诉讼的崭新发展阶段。

In the new round of judicial reform promoted by the application of artificial intelli- gence （AI） , machine cognizing cases through deep learning is the premise and weakness of judicial application of AI. The practice of local courts has not yet formed an effective method of knowledge accumulation. The application of AI should be based on the construction of map of legal knowledge and the typization and factorization of referee rules. The essential facts theory is consistent with the law of AI, which can be used as the frontier theory of a neural network, word segmentation settings and knowledge map design. The specific application path is based on the essential facts theory, and the case will be deconstructed as a different level of elements by the legal experts, thus forming the elements marked large data for machine learning. The theory of essential facts is independent of the application of AI, and AI is difficult to replace judges. AI applications and civil litigation system have a mutually reinforcing relationship, and their deep integration will open up a new stage of develooment of Chinese civil litigation.