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

05/22/2020

∙

by   Xiangjue Dong, et al.

∙

0

∙

share

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 F1-scores of 79.0 becoming one of the highest performing systems among 36 participants in this shared task.