Revisiting Contextual Toxicity Detection in Conversations
Revisiting Contextual Toxicity Detection in Conversations
复制标题
DOI：
10.1145/3561390
复制
发表时间：
2021-11
影响因子：
2.1
通讯作者：
Julia Ive;Atijit Anuchitanukul;Lucia Specia
中科院分区：
文献类型：
--
作者：
Julia Ive;Atijit Anuchitanukul;Lucia Specia
了解用户对话中的毒性无疑是一个重要问题。解决“隐蔽”或隐含的毒性案例特别困难，需要背景信息。之前很少有研究分析对话上下文对人类感知或自动检测模型的影响。我们更深入地研究这两个方向。我们首先分析现有的上下文数据集，发现人类的毒性标签通常受到对话结构、极性和上下文主题的影响。然后，我们建议通过引入和评估（a）用于上下文毒性检测的神经架构（了解对话结构）和（b）可以帮助建模上下文毒性检测的数据增强策略，将这些发现引入计算检测模型中。我们的结果表明，了解对话结构的神经架构具有令人鼓舞的潜力。我们还证明，此类模型可以从合成数据中受益，尤其是在社交媒体领域。
Understanding toxicity in user conversations is undoubtedly an important problem. Addressing “covert” or implicit cases of toxicity is particularly hard and requires context. Very few previous studies have analysed the influence of conversational context in human perception or in automated detection models. We dive deeper into both these directions. We start by analysing existing contextual datasets and find that toxicity labelling by humans is in general influenced by the conversational structure, polarity, and topic of the context. We then propose to bring these findings into computational detection models by introducing and evaluating (a) neural architectures for contextual toxicity detection that are aware of the conversational structure, and (b) data augmentation strategies that can help model contextual toxicity detection. Our results show the encouraging potential of neural architectures that are aware of the conversation structure. We also demonstrate that such models can benefit from synthetic data, especially in the social media domain.