base_model:
- google-bert/bert-base-chinese
datasets:
- china-ai-law-challenge/cail2018
language:
- zh
license: mit
metrics:
- accuracy
- f1
pipeline_tag: text-classification
library_name: transformers
⚖️ LegalDuet: Learning Effective Representations for Legal Judgment Prediction
This model, presented in the paper LegalDuet: Learning Effective Representations for Legal Judgment Prediction through a Dual-View Legal Clue Reasoning, enhances the accuracy of Legal Judgment Prediction (LJP) by employing a dual-view legal reasoning mechanism. This mechanism combines Law Case Reasoning (using past legal decisions) and Legal Ground Reasoning (matching legal clues between criminal cases and legal decisions) to create a tailored embedding space for improved legal judgment prediction. LegalDuet achieves state-of-the-art performance on the CAIL2018 dataset.
The model is based on google-bert/bert-base-chinese and can be loaded using the Transformers library.
Code: https://github.com/NEUIR/LegalDuet Project Page: http://cail.cipsc.org.cn/task_summit.html?raceID=1&cail_tag=2018