JurisCTC: Enhancing Legal Judgment Prediction via Cross-Domain Transfer and Contrastive Learning

2025-04-24

AI Summary

Legal judgment prediction (LJP) faces challenges including difficulty in cross-domain knowledge transfer between civil and criminal law, scarcity of labeled data, and weak modeling capacity for long legal texts. To address these, this paper proposes a legal-domain adaptive framework integrating cross-domain transfer learning and contrastive learning. Methodologically, it is the first to jointly leverage unsupervised domain adaptation (UDA) and cross-domain contrastive learning to collaboratively optimize pretrained language models, enabling discriminative feature transfer without cross-domain labeled data. Experiments show that our approach achieves 78.83% accuracy on LJP—significantly outperforming both the strongest baseline (76.59%) and existing large language model–based alternatives. Its primary contributions are: (1) establishing the first cross-domain contrastive transfer paradigm tailored to the legal vertical domain; (2) effectively mitigating domain shift and data sparsity; and (3) providing a novel pathway for low-resource legal NLP.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Natural Language Processing: Safety and Robustness
• Application Domains: Humanities & Computational Social Science

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In recent years, Unsupervised Domain Adaptation (UDA) has gained significant attention in the field of Natural Language Processing (NLP) owing to its ability to enhance model generalization across diverse domains. However, its application for knowledge transfer between distinct legal domains remains largely unexplored. To address the challenges posed by lengthy and complex legal texts and the limited availability of large-scale annotated datasets, we propose JurisCTC, a novel model designed to improve the accuracy of Legal Judgment Prediction (LJP) tasks. Unlike existing approaches, JurisCTC facilitates effective knowledge transfer across various legal domains and employs contrastive learning to distinguish samples from different domains. Specifically, for the LJP task, we enable knowledge transfer between civil and criminal law domains. Compared to other models and specific large language models (LLMs), JurisCTC demonstrates notable advancements, achieving peak accuracies of 76.59% and 78.83%, respectively.

Problem

• Enhancing legal judgment prediction accuracy across domains
• Addressing lengthy legal texts and limited annotated data
• Transferring knowledge between civil and criminal law

Innovation

• Cross-domain transfer for legal judgment prediction
• Contrastive learning to distinguish legal domains
• Enhancing accuracy in civil and criminal law