CofiPara: A Coarse-to-fine Paradigm for Multimodal Sarcasm Target Identification with Large Multimodal Models

arXiv:2405.00390v1 Announce Type: new
Abstract: Social media abounds with multimodal sarcasm, and identifying sarcasm targets is particularly challenging due to the implicit incongruity not directly evident in the text and image modalities. Current methods for Multimodal Sarcasm Target Identification (MSTI) predominantly focus on superficial indicators in an end-to-end manner, overlooking the nuanced understanding of multimodal sarcasm conveyed through both the text and image. This paper proposes a versatile MSTI framework with a coarse-to-fine paradigm, by augmenting sarcasm explainability with reasoning …

abstract arxiv cs.cl current focus identification image large multimodal models media multimodal multimodal models paradigm sarcasm social social media targets text type