Toxicity Detection can be Sensitive to the Conversational Context
This addresses the issue of context-dependent toxicity detection for online moderation, which is incremental as it builds on existing datasets and methods.
The paper tackles the problem of toxicity detection being insensitive to conversational context by constructing a dataset of 10,000 posts with context-dependent labels and introducing a new task for context sensitivity estimation. It shows that practical classifiers can be developed for this task, with data augmentation via knowledge distillation further improving performance.
User posts whose perceived toxicity depends on the conversational context are rare in current toxicity detection datasets. Hence, toxicity detectors trained on existing datasets will also tend to disregard context, making the detection of context-sensitive toxicity harder when it does occur. We construct and publicly release a dataset of 10,000 posts with two kinds of toxicity labels: (i) annotators considered each post with the previous one as context; and (ii) annotators had no additional context. Based on this, we introduce a new task, context sensitivity estimation, which aims to identify posts whose perceived toxicity changes if the context (previous post) is also considered. We then evaluate machine learning systems on this task, showing that classifiers of practical quality can be developed, and we show that data augmentation with knowledge distillation can improve the performance further. Such systems could be used to enhance toxicity detection datasets with more context-dependent posts, or to suggest when moderators should consider the parent posts, which often may be unnecessary and may otherwise introduce significant additional cost.