Legal Judgment Prediction and Explanation Extraction

Abstract

Over the past few years, there has been an increase in interest in predicting court decisions. On the one hand, as society has continued to expand, numerous social ties and social contradictions have grown more complex, and the number of court cases has rapidly increased, adding to the difficulty of convicting and sentencing case management. To get a thorough judgement basis, judges and pertinent case-handling staff generally need to manually review a substantial number of materials and legal documents. This technique requires a lot of time and labor, and it is not very productive. An automated system that could assist a judge in predicting the outcome of a case would help expedite the judicial process. For such a system to be practically useful, predictions by the system should be explainable. To promote research in developing such a system, we introduce LJPE (Legal Judgment Prediction and Explanation) Dataset for the Pakistan Legal Documents. LJPE is a large corpus of 11k Pakistan Supreme Court cases annotated with original court decisions. A portion of the corpus (a separate test set) is annotated with gold standard explanations by legal experts. We experiment with a battery of baseline models for case predictions and propose a hierarchical occlusion-based model for explainability.