CFMS: Towards Explainable and Fine-Grained Chinese Multimodal Sarcasm Detection Benchmark

TL;DR

This paper introduces CFMS, a fine-grained Chinese multimodal sarcasm dataset with annotations for sarcasm, target, and explanation, and proposes a reinforcement learning strategy to improve sarcasm detection models.

Contribution

It creates the first fine-grained Chinese multimodal sarcasm dataset with detailed annotations and introduces a novel RL-augmented in-context learning method for better sarcasm detection.

Findings

Fine-grained annotations improve sarcasm explanation generation.

The parallel Chinese-English metaphor subset reveals model limitations in metaphoric reasoning.

The proposed PGDS method outperforms existing baselines in key tasks.

Abstract

Multimodal sarcasm detection has recently garnered significant attention. However, existing benchmarks suffer from coarse-grained annotations and limited cultural coverage, which hinder research into fine-grained semantic understanding. To address this, we construct CFMS, the first fine-grained multimodal sarcasm dataset tailored for Chinese social media. It comprises 2,796 high-quality image-text pairs and provides a triple-level annotation framework: sarcasm identification, target recognition, and explanation generation. We find that the fine-grained explanation annotations effectively guide AI in generating images with explicit sarcastic intent. Furthermore, we curate a high-consistency parallel Chinese-English metaphor subset (200 entries each), revealing significant limitations of current models in metaphoric reasoning. To overcome the constraints of traditional retrieval methods, we propose a Reinforcement Learning-augmented In-Context Learning strategy (PGDS) to dynamically optimize exemplar selection. Extensive experiments demonstrate that CFMS provides a solid foundation for building reliable multimodal sarcasm understanding systems, and the PGDS method significantly outperforms existing baselines on key tasks. Our data and code are available at https://anonymous.4open.science/r/CFMS-E8F9.