群体自适应阈值优化用于鲁棒的AI生成文本检测
Group-Adaptive Threshold Optimization for Robust AI-Generated Text Detection
AI总结：
本文提出FairOPT算法，通过群体特定的阈值优化减少AI生成文本检测中的群体差异，提升分类鲁棒性。
摘要：
大语言模型（LLMs）的发展使得区分人类撰写文本与AI生成文本变得困难。为应对这一挑战，已开发出几种AI文本检测器，这些检测器通常使用固定的全局阈值（例如θ=0.5）进行分类。然而，一个通用的阈值可能无法考虑到子群体的分布差异。例如，当使用固定阈值时，检测器在较短的人类撰写文本上会产生更多的假阳性错误，在较长文本中对神经质的写作风格会产生更多的阳性分类。这些差异可能导致对某些群体的误分类。我们通过引入FairOPT算法，针对概率AI文本检测器进行群体特定的阈值优化。我们根据属性（如文本长度和写作风格）将数据划分为子群体，并实施FairOPT以学习每个群体的决策阈值以减少差异。FairOPT在九个检测器和三个异质数据集中展示了显著的差异缓解效果，并通过减少整体差异27.4%（在五个指标上）来显著缓解最坏情况问题，同时仅略微牺牲了准确性（0.005%）。我们的框架为通过后处理实现更稳健的AI生成内容检测铺平了道路。我们发布我们的数据、代码和项目信息在URL。
英文摘要：
The advancement of large language models (LLMs) has made it difficult to differentiate human-written text from AI-generated text. Several AI-text detectors have been developed in response, which typically utilize a fixed global threshold (e.g., $θ= 0.5$) to classify machine-generated text. However, one universal threshold could fail to account for distributional variations by subgroups. For example, when using a fixed threshold, detectors make more false positive errors on shorter human-written text, and more positive classifications of neurotic writing styles among long texts. These discrepancies can lead to misclassifications that disproportionately affect certain groups. We address this critical limitation by introducing FairOPT, an algorithm for group-specific threshold optimization for probabilistic AI-text detectors. We partitioned data into subgroups based on attributes (e.g., text length and writing style) and implemented FairOPT to learn decision thresholds for each group to reduce discrepancy. FairOPT showed notable discrepancy mitigation across nine detectors and three heterogeneous datasets, and the remarkable mitigation of the minimax problem by decreasing overall discrepancy 27.4% across five metrics while minimally sacrificing accuracy by 0.005%. Our framework paves the way for more robust classification in AI-generated content detection via post-processing. We release our data, code, and project information at URL.
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。
