CASCADE: Contextual Sarcasm Detection in Online Discussion Forums

TL;DR

CASCADE introduces a hybrid model for sarcasm detection that combines content analysis with contextual and user-specific information, significantly improving accuracy on social media data.

Contribution

The paper presents a novel hybrid approach integrating contextual discourse and user embeddings for sarcasm detection in online forums.

Findings

Enhanced sarcasm detection accuracy on Reddit data

Effective use of user embeddings for personalized context understanding

Hybrid model outperforms content-only baselines

Abstract

The literature in automated sarcasm detection has mainly focused on lexical, syntactic and semantic-level analysis of text. However, a sarcastic sentence can be expressed with contextual presumptions, background and commonsense knowledge. In this paper, we propose CASCADE (a ContextuAl SarCasm DEtector) that adopts a hybrid approach of both content and context-driven modeling for sarcasm detection in online social media discussions. For the latter, CASCADE aims at extracting contextual information from the discourse of a discussion thread. Also, since the sarcastic nature and form of expression can vary from person to person, CASCADE utilizes user embeddings that encode stylometric and personality features of the users. When used along with content-based feature extractors such as Convolutional Neural Networks (CNNs), we see a significant boost in the classification performance on a large Reddit corpus.