Legal knowledge management for prosecutors based on judgment prediction and error analysis from indictments

Legal AI aims to provide improved knowledge management services based on legal documents. Existing legal judgment prediction datasets mainly use court verdicts. However, for prosecutors, the use of indictments for judgment predictions can help detecting inconsistencies between predictions and prosecution, providing prosecutors with more accurate references to laws and charges through error analysis. In this study, we collect a dataset called TWLJP, which contains 342,754 indictments. We compared three possible messaging passing architectures among the law, regulation, and accusation cause prediction tasks, i.e. independent, topological, and interactive. The result shows that interactive message passing among the three tasks achieved the best Macro-F1 performance of 95.2 %, 79.62 %, and 65.84 % for laws, regulations, and accusation cause prediction, respectively. We further improve the prediction of accusation cause from 8.8 % macro-F1 to 62.3 % for underperformed accusation causes via Prompt-Based Learning. Finally, in view of the situation where the charge prediction are written in various ways, we adopted a lenient approach to assess the accusation and improved the accusation performance to 77.2 %.

Automated summary

This study aims to provide improved knowledge management services based on legal documents. By using indictments for judgment predictions, inconsistencies between predictions and prosecution can be detected, providing prosecutors with more accurate references to laws and charges. The study compared different messaging passing architectures and achieved the best performance with interactive message passing. The prediction accuracy of accusation causes was further improved through Prompt-Based Learning.