LePREC: Reasoning as Classification over Structured Factors for Assessing Relevance of Legal Issues
Fanyu Wang, Xiaoxi Kang, Paul Burgess, Aashish Srivastava, Chetan Arora, Adnan Trakic, Lay-Ki Soon, Md Khalid Hossain, Lizhen Qu
Why It Matters
What makes this one worth your time
This research addresses a critical gap in legal resource accessibility by improving the precision of legal issue identification, which can aid in better legal decision-making.
LePREC enhances legal issue identification by integrating neural and symbolic reasoning methods.
Summary
The paper presents LePREC, a neuro-symbolic framework that combines neural generation with structured statistical reasoning to improve the identification of legal issues from court cases, achieving a notable performance improvement over existing LLMs.
Key contributions
- Development of the LePREC framework that integrates neural and symbolic components for legal reasoning.
- Creation of a dataset from real-world Malaysian court cases annotated by legal experts.
- Demonstration of a 30-40% performance improvement over advanced LLM baselines.
Notable insights
- The combination of neural generation with structured statistical reasoning allows for both interpretability and data efficiency, which is often lacking in purely neural approaches.
- The use of correlation-based statistical classification to enhance relevance decisions is a novel approach in the legal domain.
Possible limitations
- The abstract does not discuss the generalizability of the framework beyond the Malaysian legal context.
- Potential limitations in the dataset size and diversity may affect the robustness of the findings.
- Not stated in the abstract.
Abstract
arXiv:2604.19464v1 Announce Type: cross Abstract: More than half of the global population struggles to meet their civil justice needs due to limited legal resources. While Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, significant challenges remain even at the foundational step of legal issue identification. To investigate LLMs' capabilities in this task, we constructed a dataset from 769 real-world Malaysian Contract Act court cases, using GPT-4o to extract facts and generate candidate legal issues, annotated by senior legal experts, which reveals a critical limitation: while LLMs generate diverse issue candidates, their precision remains inadequate (GPT-4o achieves only 62%). To address this gap, we propose LePREC (Legal Professional-inspired Reasoning Elicitation and Classification), a neuro-symbolic framework combining neural generation with structured statistical reasoning. LePREC consists of: (1) a neuro component leverages LLMs to transform legal descriptions into question-answer pairs representing diverse analytical factors, and (2) a symbolic component applies sparse linear models over these discrete features, learning explicit algebraic weights that identify the most informative reasoning factors. Unlike end-to-end neural approaches, LePREC achieves interpretability through transparent feature weighting while maintaining data efficiency through correlation-based statistical classification. Experiments show a 30-40% improvement over advanced LLM baselines, including GPT-4o and Claude, confirming that correlation-based factor-issue analysis offers a more data-efficient solution for relevance decisions.