Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Beyond the Literal: Decomposing Pragmatic Intent in Multimodal Meme Understanding

About

When asked what a meme or sarcastic post means, Large Vision Language Models (LVLMs) tend to describe what the image shows rather than what the author is trying to communicate. Standard instruction tuning entangles a post's literal content with its pragmatic meaning, letting surface-level details contaminate the final response. We reframe meme understanding as a problem of literal-pragmatic decomposition and propose \textbf{Intent Projection}, a framework that separates the two signals at the representation, output, and objective levels within a single LVLM backbone. At the representation level, an orthogonal projection module removes dominant unimodal directions from the fused image-text representation, retaining only the pragmatic residual, while a surface-real affect classifier anchors the decoder with a discrete tag that names the polarity gap. At the output level, the model externalizes a structured reasoning chain, and at the objective level a contrastive reward explicitly penalizes answers that restate the literal description. Across six multimodal benchmarks, Intent Projection consistently outperforms open-source baselines and narrows the gap to proprietary models, with the largest gains on high-divergence posts where literal collapse is most damaging.

Zhengyi Zhao, Shubo Zhang, Zezhong Wang, Luyao Ye, Huimin Wang, Hanqi Yan, Binyang Li, Kam-Fai Wong, Yulan He• 2026

Related benchmarks

TaskDatasetResultRank
Meme GenerationMemeReaCon PC-G (test)
BERTScore F159.8
21
Meme GenerationMemeReaCon PI-G (test)
BERTScore F141.3
21
Meme GenerationMemeCap (test)
BERTScore-F154.5
21
Meme GenerationMET-Meme (test)
BERTScore F151.7
21
Meme GenerationMUSE (test)
BERTScore F158
21
Meme GenerationGOAT-G (test)
BERTScore F153.1
21
Meme ClassificationGOAT-C (test)
Accuracy79.4
12
Meme ClassificationMemeReaCon CMI-C (test)
Accuracy79.6
12
Meme ClassificationMemeReaCon CSAC-C (test)
Accuracy65
12
Meme ClassificationMMSD 2.0 (test)
Accuracy84.8
12
Showing 10 of 10 rows

Other info

Follow for update