Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Was that Sarcasm?: A Literature Survey on Sarcasm Detection

Harleen Kaur Bagga, Jasmine Bernard, Sahil Shaheen, Sarthak Arora

TL;DR

This survey investigates sarcasm detection in NLP, focusing on the challenge of interpreting sarcasm and the role of datasets in benchmarking. It categorizes approaches into linguistic/context-based, embeddings/topic modeling, multi-modal, and graph-based methods, and highlights key datasets such as MUStARD and SARC. It reviews influential models including MIARN, term-weighted neural language models, Pan's intra-/inter-modal incongruity framework, and KnowleNet, along with prompting strategies like SarcPrompt. The findings show that combining contextual cues, affective representations, and cross-modal information yields substantial gains, and point to future work with large language models, multimodal fusion, multilingual data, synthetic data, and metaphor detection to advance sarcasm detection.

Abstract

Sarcasm is hard to interpret as human beings. Being able to interpret sarcasm is often termed as a sign of intelligence, given the complex nature of sarcasm. Hence, this is a field of Natural Language Processing which is still complex for computers to decipher. This Literature Survey delves into different aspects of sarcasm detection, to create an understanding of the underlying problems faced during detection, approaches used to solve this problem, and different forms of available datasets for sarcasm detection.

Was that Sarcasm?: A Literature Survey on Sarcasm Detection

TL;DR

This survey investigates sarcasm detection in NLP, focusing on the challenge of interpreting sarcasm and the role of datasets in benchmarking. It categorizes approaches into linguistic/context-based, embeddings/topic modeling, multi-modal, and graph-based methods, and highlights key datasets such as MUStARD and SARC. It reviews influential models including MIARN, term-weighted neural language models, Pan's intra-/inter-modal incongruity framework, and KnowleNet, along with prompting strategies like SarcPrompt. The findings show that combining contextual cues, affective representations, and cross-modal information yields substantial gains, and point to future work with large language models, multimodal fusion, multilingual data, synthetic data, and metaphor detection to advance sarcasm detection.

Abstract

Sarcasm is hard to interpret as human beings. Being able to interpret sarcasm is often termed as a sign of intelligence, given the complex nature of sarcasm. Hence, this is a field of Natural Language Processing which is still complex for computers to decipher. This Literature Survey delves into different aspects of sarcasm detection, to create an understanding of the underlying problems faced during detection, approaches used to solve this problem, and different forms of available datasets for sarcasm detection.

Paper Structure

This paper contains 23 sections, 7 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Linguistic and Context-based Approaches
  4. Word Embeddings and Topic Modeling
  5. Multi-modal Approaches
  6. Graph-based Approaches
  7. Sarcasm Datasets
  8. Data Sources
  9. Data Characteristics and Preliminary Analysis
  10. MUStARD (2019)
  11. SARC (2018)
  12. Embeddings
  13. Topic-enriched (2019)
  14. Affective Representations (2018)
  15. Approaches
  16. ...and 8 more sections

Figures (7)

  • Figure 1: A sarcastic utterance and its context from the dataset represented by video frames and transcript according to Castro et al. castro
  • Figure 2: Character-label ratio per source according to Castro et al. castro
  • Figure 3: An overview of the AWES framework according to Agrawal et al. ameeta
  • Figure 4: Accuracy over the five different feature sets according to Bamman et al.bamman
  • Figure 5: How the model looks at multi-modality in sarcasm according to Pan et al. pan
  • ...and 2 more figures