Deep and Dense Sarcasm Detection

TL;DR

This paper introduces a deep 56-layer dense network for sarcasm detection that effectively models textual features without relying on external context, achieving competitive results and better classifying sarcastic utterances.

Contribution

It presents a novel deep dense connectivity model for sarcasm detection that outperforms shallow models and state-of-the-art architectures using only local text features.

Findings

Competitive results with state-of-the-art models

Effective detection of sarcastic utterances without external context

Deeper understanding of feature map generation

Abstract

Recent work in automated sarcasm detection has placed a heavy focus on context and meta-data. Whilst certain utterances indeed require background knowledge and commonsense reasoning, previous works have only explored shallow models for capturing the lexical, syntactic and semantic cues present within a text. In this paper, we propose a deep 56 layer network, implemented with dense connectivity to model the isolated utterance and extract richer features therein. We compare our approach against recent state-of-the-art architectures which make considerable use of extrinsic information, and demonstrate competitive results whilst using only the local features of the text. Further, we provide an analysis of the dependency of prior convolution outputs in generating the final feature maps. Finally a case study is presented, supporting that our approach accurately classifies additional uses of clear sarcasm, which a standard CNN misclassifies.