不要被愚弄：面向用户意图对抗攻击的冒犯性语言检测中的池化策略
Don't be a Fool: Pooling Strategies in Offensive Language Detection from User-Intended Adversarial Attacks
AI总结：
针对用户利用文本噪声规避过滤的用户意图对抗攻击，提出逐层池化防御策略，关注前层以同时捕获冒犯性和token嵌入，无需专门训练即提升鲁棒性，使干净文本预训练模型达到噪声预训练模型的检测效果。
摘要：
冒犯性语言检测是过滤辱骂表达和改善在线用户体验的重要任务。然而，恶意用户常试图通过引入文本噪声来规避过滤系统。本文将这些规避行为提出为用户意图的对抗攻击，即插入特殊符号或利用韩语的独特特征。此外，我们引入了简单而有效的逐层池化策略来防御所提出的攻击，重点关注前面的层而不仅是最后一层，以同时捕获冒犯性和token嵌入。我们证明，即使攻击率增加，这些池化策略对性能下降也更具鲁棒性，而无需直接训练此类模式。值得注意的是，我们发现通过采用这些池化策略，在干净文本上预训练的模型在检测受攻击的冒犯性语言时，能够取得与在噪声文本上预训练的模型相当的性能。
英文摘要：
Offensive language detection is an important task for filtering out abusive expressions and improving online user experiences. However, malicious users often attempt to avoid filtering systems through the involvement of textual noises. In this paper, we propose these evasions as user-intended adversarial attacks that insert special symbols or leverage the distinctive features of the Korean language. Furthermore, we introduce simple yet effective pooling strategies in a layer-wise manner to defend against the proposed attacks, focusing on the preceding layers not just the last layer to capture both offensiveness and token embeddings. We demonstrate that these pooling strategies are more robust to performance degradation even when the attack rate is increased, without directly training of such patterns. Notably, we found that models pre-trained on clean texts could achieve a comparable performance in detecting attacked offensive language, to models pre-trained on noisy texts by employing these pooling strategies.
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。
