Sarcasm Detection
Sarcasm detection, the task of automatically identifying sarcastic intent in text and multimodal data, aims to improve natural language understanding and sentiment analysis. Current research focuses on leveraging large language models (LLMs) and multimodal models, often incorporating contextual cues like prosody, visual information, and emoji, to overcome the challenges posed by sarcasm's inherent ambiguity. These advancements are significant for applications ranging from social media monitoring to improved human-computer interaction, as accurate sarcasm detection enhances the ability of systems to understand nuanced human communication. However, challenges remain in generalizing across different datasets and sarcasm styles, highlighting the need for more robust and explainable models.
Papers
TextMI: Textualize Multimodal Information for Integrating Non-verbal Cues in Pre-trained Language Models
Md Kamrul Hasan, Md Saiful Islam, Sangwu Lee, Wasifur Rahman, Iftekhar Naim, Mohammed Ibrahim Khan, Ehsan Hoque
Borrowing Human Senses: Comment-Aware Self-Training for Social Media Multimodal Classification
Chunpu Xu, Jing Li