Confronting LLMs with Traditional ML: Rethinking the Fairness of Large Language Models in Tabular Classifications
作者: Yanchen Liu, Srishti Gautam, Jiaqi Ma, Himabindu Lakkaraju
分类: cs.CL, cs.LG
发布日期: 2023-10-23 (更新: 2024-04-02)
备注: NAACL 2024 (Main Conference)
💡 一句话要点
探讨大型语言模型在表格分类中的公平性问题
🎯 匹配领域: 支柱九：具身大模型 (Embodied Foundation Models)
关键词: 大型语言模型 表格分类 社会偏见 公平性 偏见缓解 实验研究 机器学习 人工智能
📋 核心要点
- 现有大型语言模型在表格分类任务中存在显著的社会偏见，影响其公平性。
- 通过实验，论文探讨了LLMs在分类时的信息来源及其社会偏见的影响。
- 研究表明，尽管偏见缓解方法有所成效，但LLMs的公平性仍低于传统机器学习模型。
📝 摘要（中文）
近年来的文献表明，大型语言模型（LLMs）在表格任务分类中具有潜力。然而，LLMs显示出有害的社会偏见，反映了社会中的刻板印象和不平等。本文通过一系列实验探讨LLMs在表格任务分类中的信息来源、社会偏见的影响及其对公平性的后果。研究发现，LLMs从训练数据中继承社会偏见，显著影响其在表格分类任务中的公平性。此外，尽管上下文学习和微调对偏见缓解有一定效果，但不同子组之间的公平性指标差距仍大于传统机器学习模型，如随机森林和浅层神经网络。我们还展示了上下文示例的标签翻转可以显著减少偏见，进一步强调了LLMs内在偏见的存在。
🔬 方法详解
问题定义：本文旨在解决大型语言模型在表格分类任务中存在的社会偏见问题。现有方法未能有效缓解这些偏见，导致分类结果的不公平性。
核心思路：论文通过实验分析LLMs在表格任务中的分类决策过程，探讨其如何受到训练数据中社会偏见的影响，并提出标签翻转作为偏见缓解的策略。
技术框架：研究采用实验设计，首先分析LLMs的分类决策过程，然后通过对比不同模型（如随机森林和浅层神经网络）评估公平性指标，最后实施标签翻转策略进行偏见缓解。
关键创新：最重要的创新在于揭示了LLMs的社会偏见不仅源于下游任务数据集，还源于其预训练语料库，强调了偏见的内在性。
关键设计：实验中采用了多种公平性指标进行评估，并通过上下文学习和微调进行偏见缓解，此外，标签翻转策略被证明能有效减少偏见。
🖼️ 关键图片
📊 实验亮点
实验结果显示，LLMs在表格分类任务中的公平性指标差距明显大于传统机器学习模型，且通过标签翻转策略，偏见显著减少。这一发现强调了LLMs内在偏见的严重性，且在偏见缓解方面的效果有限。
🎯 应用场景
该研究的潜在应用领域包括金融、医疗和招聘等高风险行业，这些领域普遍使用表格数据进行决策。通过提高LLMs在这些领域的公平性，可以减少因社会偏见导致的不平等待遇，提升决策的公正性和透明度。未来，研究成果有望推动更公平的人工智能系统的开发与应用。
📄 摘要（原文）
Recent literature has suggested the potential of using large language models (LLMs) to make classifications for tabular tasks. However, LLMs have been shown to exhibit harmful social biases that reflect the stereotypes and inequalities present in society. To this end, as well as the widespread use of tabular data in many high-stake applications, it is important to explore the following questions: what sources of information do LLMs draw upon when making classifications for tabular tasks; whether and to what extent are LLM classifications for tabular data influenced by social biases and stereotypes; and what are the consequential implications for fairness? Through a series of experiments, we delve into these questions and show that LLMs tend to inherit social biases from their training data which significantly impact their fairness in tabular classification tasks. Furthermore, our investigations show that in the context of bias mitigation, though in-context learning and finetuning have a moderate effect, the fairness metric gap between different subgroups is still larger than that in traditional machine learning models, such as Random Forest and shallow Neural Networks. This observation emphasizes that the social biases are inherent within the LLMs themselves and inherited from their pretraining corpus, not only from the downstream task datasets. Besides, we demonstrate that label-flipping of in-context examples can significantly reduce biases, further highlighting the presence of inherent bias within LLMs.