Transformer-based Context-aware Sarcasm Detection in Conversation Threads from Social Media

FIGURATIVE LANGUAGE PROCESSING(2020)

引用 4|浏览54
暂无评分
摘要
We present a transformer-based sarcasm detection model that accounts for the context from the entire conversation thread for more robust predictions. Our model uses deep transformer layers to perform multi-head attentions among the target utterance and the relevant context in the thread. The context-aware models are evaluated on two datasets from social media, Twitter and Reddit, and show 3.1% and 7.0% improvements over their baselines. Our best models give the Fl-scores of 79.0% and 75.0% for the Twitter and Reddit datasets respectively, becoming one of the highest performing systems among 36 participants in this shared task.
更多
查看译文
关键词
sarcasm,conversation,transformer-based,context-aware
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要