通过分析模型表示的坍缩现象，提出提升语言模型公平性的新方法。
Collapsed Language Models Promote Fairness
- 基于神经坍缩现象设计公平性优化策略
- 在多个数据集上显著降低性别/种族偏见，性能不降
- 适合需要公平性的NLP应用开发与研究
为缓解近期成功预训练语言模型中隐含的社会偏见，学界提出了多种促进模型公平性的方法，包括提示工程、数据增强和正则化微调等。尽管如此，要建立对公平性的原则性理解并设计出能持续去偏的算法仍具挑战。本文通过严格评估神经坍缩——深度网络最后一层表征与分类器中的学习现象——在与公平性相关词汇上的表现，发现去偏语言模型在词元表征与词嵌入之间表现出坍缩对齐。更重要的是，这一观察启发我们设计一种有原则的微调方法，可在广泛去偏方法中有效提升公平性，同时保持语言模型在标准自然语言理解任务上的性能。代码已公开于 https://github.com/Xujxyang/Fairness-NC-main。
原文摘要 · Abstract (English)
To mitigate societal biases implicitly encoded in recent successful pretrained language models, a diverse array of approaches have been proposed to encourage model fairness, focusing on prompting, data augmentation, regularized fine-tuning, and more. Despite the development, it is nontrivial to reach a principled understanding of fairness and an effective algorithm that can consistently debias language models. In this work, by rigorous evaluations of Neural Collapse -- a learning phenomenon happen in last-layer representations and classifiers in deep networks -- on fairness-related words, we find that debiased language models exhibit collapsed alignment between token representations and word embeddings. More importantly, this observation inspires us to design a principled fine-tuning method that can effectively improve fairness in a wide range of debiasing methods, while still preserving the performance of language models on standard natural language understanding tasks. We attach our code at https://github.com/Xujxyang/Fairness-NC-main.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。