@inproceedings{zeng-etal-2026-cafe,
title = "{CAFE}: Caption-Anchored Feature Enrichment and Evidence Routing for Multimodal Sarcasm Detection",
author = "Zeng, Puchu and
Luo, Yun and
li, Qianqian and
Ma, Yanzhou and
Junyi, Liu and
Fengxuan, Song and
Chonghuifang",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the {A}sia-Pacific Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-long.142/",
pages = "3103--3119",
ISBN = "979-8-89176-439-2",
abstract = "Multimodal sarcasm detection aims to identify sarcasm in image-text posts by capturing how visual content changes the interpretation of short informal text. Existing methods have advanced the task through multimodal fusion, caption-based cues, and fact/sentiment incongruity modeling, but their evidence use is often controlled inside latent feature channels or reduced discrepancy scores. This makes it difficult to decide which visual information should be selectively weighted for a particular post. We propose CAFE, a caption-anchored feature enrichment and evidence routing framework that decides how to use visual evidence for each sample under ambiguity. CAFE enriches each image with two generated descriptions, one for visible facts and one for the image-text relation, then estimates how useful each description is for the current post and fuses caption-calibrated contrast evidence with direct cross-modal evidence for prediction. Experiments show that CAFE achieves state-of-the-art performance on the primary MMSD2.0 benchmark with 87.81+/-0.33{\%} F1 (mean over 5 seeds). Component ablations on the original MMSD provide an additional evaluation."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zeng-etal-2026-cafe">
<titleInfo>
<title>CAFE: Caption-Anchored Feature Enrichment and Evidence Routing for Multimodal Sarcasm Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Puchu</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yun</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qianqian</namePart>
<namePart type="family">li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yanzhou</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Liu</namePart>
<namePart type="family">Junyi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Song</namePart>
<namePart type="family">Fengxuan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name>
<namePart>Chonghuifang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-439-2</identifier>
</relatedItem>
<abstract>Multimodal sarcasm detection aims to identify sarcasm in image-text posts by capturing how visual content changes the interpretation of short informal text. Existing methods have advanced the task through multimodal fusion, caption-based cues, and fact/sentiment incongruity modeling, but their evidence use is often controlled inside latent feature channels or reduced discrepancy scores. This makes it difficult to decide which visual information should be selectively weighted for a particular post. We propose CAFE, a caption-anchored feature enrichment and evidence routing framework that decides how to use visual evidence for each sample under ambiguity. CAFE enriches each image with two generated descriptions, one for visible facts and one for the image-text relation, then estimates how useful each description is for the current post and fuses caption-calibrated contrast evidence with direct cross-modal evidence for prediction. Experiments show that CAFE achieves state-of-the-art performance on the primary MMSD2.0 benchmark with 87.81+/-0.33% F1 (mean over 5 seeds). Component ablations on the original MMSD provide an additional evaluation.</abstract>
<identifier type="citekey">zeng-etal-2026-cafe</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-long.142/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>3103</start>
<end>3119</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CAFE: Caption-Anchored Feature Enrichment and Evidence Routing for Multimodal Sarcasm Detection
%A Zeng, Puchu
%A Luo, Yun
%A li, Qianqian
%A Ma, Yanzhou
%A Junyi, Liu
%A Fengxuan, Song
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%A Chonghuifang
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-439-2
%F zeng-etal-2026-cafe
%X Multimodal sarcasm detection aims to identify sarcasm in image-text posts by capturing how visual content changes the interpretation of short informal text. Existing methods have advanced the task through multimodal fusion, caption-based cues, and fact/sentiment incongruity modeling, but their evidence use is often controlled inside latent feature channels or reduced discrepancy scores. This makes it difficult to decide which visual information should be selectively weighted for a particular post. We propose CAFE, a caption-anchored feature enrichment and evidence routing framework that decides how to use visual evidence for each sample under ambiguity. CAFE enriches each image with two generated descriptions, one for visible facts and one for the image-text relation, then estimates how useful each description is for the current post and fuses caption-calibrated contrast evidence with direct cross-modal evidence for prediction. Experiments show that CAFE achieves state-of-the-art performance on the primary MMSD2.0 benchmark with 87.81+/-0.33% F1 (mean over 5 seeds). Component ablations on the original MMSD provide an additional evaluation.
%U https://aclanthology.org/2026.aacl-long.142/
%P 3103-3119
Markdown (Informal)
[CAFE: Caption-Anchored Feature Enrichment and Evidence Routing for Multimodal Sarcasm Detection](https://aclanthology.org/2026.aacl-long.142/) (Zeng et al., AACL-IJCNLP 2026)
ACL
- Puchu Zeng, Yun Luo, Qianqian li, Yanzhou Ma, Liu Junyi, Song Fengxuan, and Chonghuifang. 2026. CAFE: Caption-Anchored Feature Enrichment and Evidence Routing for Multimodal Sarcasm Detection. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3103–3119, Hengqin, China. Association for Computational Linguistics.