arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.13868cs.CLcs.AIcs.LG
收藏

法律面前人人平等:公平的法律判断一致性分析

Equality before the Law: Legal Judgment Consistency Analysis for Fairness

Tsinghua University(清华大学);Beijing Powerlaw Intelligent Technology Co., Ltd.(北京幂律智能科技有限公司)

Yuzhong Wang, Chaojun Xiao, Shirong Ma, Haoxi Zhong, Cunchao Tu, Tianyang Zhang, Zhiyuan Liu, Maosong Sun

摘要:

在一个法律体系中,判决的一致性被认为是公平的最重要表现之一。然而,由于在现实世界中影响量刑的事实因素的复杂性,很少有工作已经做了定量测量的判断一致性对现实世界的数据。在本文中,我们提出了一个判断不一致性的评估指标,法律不一致系数(LInCo),旨在评估按特定特征划分的数据组之间的不一致性(例如,性别、地区、种族)。我们建议模拟法官从不同的群体与法律判断预测(LJP)模型和测量的司法不一致性的判决结果的分歧LJP模型训练在不同的群体。合成数据的实验结果验证了LInCo的有效性。我们进一步使用LInCo来探讨真实案例中的不一致性,并得出以下结论:(1)法律制度中存在区域和性别不一致,但性别不一致性远远小于区域不一致性;(2)区域不一致性的水平在不同时期变化不大;(3)总体而言,司法不一致性与刑事指控的严重程度呈负相关。此外,我们使用LInCo来评估几种去偏置方法的性能,如对抗学习,发现这些机制可以有效地帮助LJP模型避免遭受数据偏差。

英文摘要:

In a legal system, judgment consistency is regarded as one of the most important manifestations of fairness. However, due to the complexity of factual elements that impact sentencing in real-world scenarios, few works have been done on quantitatively measuring judgment consistency towards real-world data. In this paper, we propose an evaluation metric for judgment inconsistency, Legal Inconsistency Coefficient (LInCo), which aims to evaluate inconsistency between data groups divided by specific features (e.g., gender, region, race). We propose to simulate judges from different groups with legal judgment prediction (LJP) models and measure the judicial inconsistency with the disagreement of the judgment results given by LJP models trained on different groups. Experimental results on the synthetic data verify the effectiveness of LInCo. We further employ LInCo to explore the inconsistency in real cases and come to the following observations: (1) Both regional and gender inconsistency exist in the legal system, but gender inconsistency is much less than regional inconsistency; (2) The level of regional inconsistency varies little across different time periods; (3) In general, judicial inconsistency is negatively correlated with the severity of the criminal charges. Besides, we use LInCo to evaluate the performance of several de-bias methods, such as adversarial learning, and find that these mechanisms can effectively help LJP models to avoid suffering from data bias.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑