Computer Science
Target-Augmented Shared Fusion-based Multimodal Sarcasm Explanation Generation
P. Goel, D. S. Chauhan, et al.
Sarcasm often hides its target — and TURBO uncovers it. This study introduces TURBO, a Target-aUgmented shaRed fusion-Based sarcasm explanation model that leverages a novel shared-fusion mechanism to fuse image and caption and explicitly models the sarcasm target to generate clearer explanations. Evaluated on MORE+, TURBO outperforms baselines by an average +3.3%, and human assessments find its explanations superior; the authors also examine LLMs in zero/one-shot settings and note their limitations. This research was conducted by Palaash Goel, Dushyant Singh Chauhan, and Md Shad Akhtar.
Related Publications
Explore these studies to deepen your understanding of the subject.
