Interpretable Charge Predictions for Criminal Cases: Learning to Generate Court Views from Fact Descriptions

TL;DR

This paper introduces a label-conditioned sequence-to-sequence model for generating interpretable court views from criminal case fact descriptions, enhancing charge prediction interpretability and legal document automation.

Contribution

It proposes a novel label-conditioned Seq2Seq approach that improves court view generation by incorporating charge labels, addressing non-distinction issues in fact descriptions.

Findings

The model effectively generates charge-discriminative court views.

Charge labels significantly improve generation quality.

Experimental results demonstrate the method's superiority.

Abstract

In this paper, we propose to study the problem of COURT VIEW GENeration from the fact description in a criminal case. The task aims to improve the interpretability of charge prediction systems and help automatic legal document generation. We formulate this task as a text-to-text natural language generation (NLG) problem. Sequenceto-sequence model has achieved cutting-edge performances in many NLG tasks. However, due to the non-distinctions of fact descriptions, it is hard for Seq2Seq model to generate charge-discriminative court views. In this work, we explore charge labels to tackle this issue. We propose a label-conditioned Seq2Seq model with attention for this problem, to decode court views conditioned on encoded charge labels. Experimental results show the effectiveness of our method.