Xiezhi: An Intelligent Judical Trial System

Nowadays, the growing integration of artificial intelligence with various fields has promoted the rapid development of each field. However, in the judicial field, the application of artificial intelligence is limited, and there is still a great demand and potential for development. In this paper, we devise an intelligent judicial trial system-XieZhi, which realizes the following functions: crime prediction, sentence prediction, relevant law articles recommendation and similar cases recommendation. Specifically, we use the BERT model to design a text classification network for the crime and sentence prediction. For the relevant law articles recommendation we create a statistical-based method, and for the similar cases recommendation, we select the approach that using TF-IDF to calculate text similarity. Through experiments, it shows that all functions of our Xiezhi system have a high accuracy and good performance.