KID: Knowledge-Injected Dual-Head Learning for Knowledge-Grounded Harmful Meme Detection
School of Economics and Management, Beijing University of Posts and Telecommunications(经济管理学院,北京邮电大学);College of Computing, Beijing Information Science and Technology University(计算机学院,北京信息科技大学)
Internet memes have become pervasive carriers of digital culture on social platforms. However, their heavy reliance on metaphors and sociocultural context also makes them subtle vehicles for harmful content, posing significant challenges for automated content moderation. Existing approaches primarily focus on intra-modal and inter-modal signal analysis, while the understanding of implicit toxicity often depends on background knowledge that is not explicitly present in the meme itself. To address this challenge, we propose KID, a Knowledge-Injected Dual-Head Learning framework for knowledge-grounded harmful meme detection. KID adopts a label-constrained distillation paradigm to decompose complex meme understanding into structured reasoning chains that explicitly link visual evidence, background knowledge, and classification labels. These chains guide the learning process by grounding external knowledge in meme-specific contexts. In addition, KID employs a dual-head architecture that jointly optimizes semantic generation and classification objectives, enabling aligned linguistic reasoning while maintaining stable decision boundaries. Extensive experiments on five multilingual datasets spanning English, Chinese, and low-resource Bengali demonstrate that KID achieves SOTA performance on both binary and multi-label harmful meme detection tasks, improving over previous best methods by 2.1%--19.7% across primary evaluation metrics. Ablation studies further confirm the effectiveness of knowledge injection and dual-head joint learning, highlighting their complementary contributions to robust and generalizable meme understanding. The code and data are available at https://github.com/PotatoDog1669/KID.
Towards Comprehensive Detection of Chinese Harmful Memes
School of Computer Science and Technology(计算机科学与技术学院);School of Foreign Languages(外国语学院);Key Laboratory of Social Computing and Cognitive Intelligence(社会计算与认知智能重点实验室);Dalian University of Technology(大连理工大学)
Distinguishing Right from Wrong in Debates: Attribution Analysis of Chinese Harmful Memes
Key Laboratory of Social Computing and Cognitive Intelligence, Dalian University of Technology(社会计算与认知智能重点实验室,大连理工大学);Singapore University of Technology and Design(新加坡科技设计大学);Data Intelligence Lab, Department of Computer Science, City University of Hong Kong(数据智能实验室,香港城市大学计算机科学系)
Detecting and Understanding Harmful Memes: A Survey
IIIT-Delhi(德里印度理工学院);Wipro AI Labs(威普罗人工智能实验室);Qatar Computing Research Institute, HBKU(哈马德·本·哈利法大学卡塔尔计算研究所);Sofia University(索非亚大学);University of Padova(帕多瓦大学);Facebook AI(脸书人工智能);Sapienza University of Rome(罗马第一大学)
MOMENTA: A Multimodal Framework for Detecting Harmful Memes and Their Targets
Johns Hopkins University(约翰斯·霍普金斯大学);Indraprastha Institute of Information Technology - Delhi(德里英迪拉普拉斯信息技术学院);Wipro AI Labs(威普罗人工智能实验室);Sofia University(索非亚大学);Qatar Computing Research Institute, HBKU(哈马德·本·哈利法大学卡塔尔计算研究所)
网络模因已经成为传播政治、心理和社会文化思想的有力手段。虽然模因通常是幽默的,但最近几天,用于拖钓、网络欺凌和虐待的有害模因不断升级。检测这样的模因是具有挑战性的,因为它们可能是高度讽刺和神秘的。此外,虽然以前的工作集中在模因的具体方面,如仇恨言论和宣传,但很少有关于一般伤害的工作。在这里,我们的目标是弥合这一差距。我们专注于两个任务:(i)检测有害的模因,(ii)识别它们所针对的社会实体。我们进一步扩展了最近发布的HarMeme数据集,该数据集涵盖了COVID-19,增加了额外的模因和一个新的主题:美国政治。为了解决这些任务,我们提出了MOMENTA(MultiModal framework for detecting harmful MemEs aNd Their tTargets),这是一种新型的多模态深度神经网络,它使用全局和局部视角来检测有害模因。MOMENTA系统地分析了输入模因的本地和全球视角(两种模式),并将其与背景语境联系起来。MOMENTA是可解释和可推广的,我们的实验表明,它优于几个强大的竞争方法。
Indraprastha Institute of Information Technology - Delhi(德里英迪拉普拉斯信息技术学院);Sofia University(索非亚大学);Wipro AI Labs(威普罗人工智能实验室);Qatar Computing Research Institute, HBKU(卡塔尔计算研究所(哈马德·本·哈利法大学))
SyRHM: Symbolic-Language-Enhanced Reasoning with Associative Retrieval for Zero-shot Harmful Meme Detection
University of California, San Diego(加利福尼亚大学圣迭戈分校);Xi’an Jiaotong-Liverpool University(西交利物浦大学);University of Washington(华盛顿大学);The Australian National University(澳大利亚国立大学);Zhejiang University(浙江大学)
DISARM: Detecting the Victims Targeted by Harmful Memes
Indraprastha Institute of Information Technology - Delhi(德里英德拉普拉斯塔信息技术学院);Wipro AI Labs(威普罗人工智能实验室);Qatar Computing Research Institute, HBKU(哈马德·本·哈利法大学卡塔尔计算研究所)
Detecting the Role of an Entity in Harmful Memes: Techniques and Their Limitations
BJIT Limited(BJIT有限公司);Qatar Computing Research Institute, HBKU(哈马德·本·哈利法大学卡塔尔计算研究所)
随着时间的推移,有害或滥用的在线内容不断增加,引起了社交媒体平台,政府机构和政策制定者的关注。这种有害或滥用的内容可能对社会产生重大的负面影响,例如,网络欺凌可能导致自杀,关于COVID-19的谣言可能导致疫苗犹豫,推广COVID-19的假治疗可能导致健康损害和死亡。在线发布和共享的内容可以是文本的、视觉的或两者的组合,例如,in a meme模因.在这里,我们描述了我们在检测有害模因中实体(英雄,恶棍,受害者)角色的实验,这是约束-2022共享任务的一部分,以及我们的任务系统。我们进一步提供了不同实验设置的比较分析(即,单峰、多峰、注意力和增强)。为了可重复性,我们公开了我们的实验代码。\url{https://github.com/robi56/harmful_memes_block_fusion}
Characterizing the Entities in Harmful Memes: Who is the Hero, the Villain, the Victim?
Indian Institute of Technology Delhi(德里印度理工学院);Wipro AI Labs(威普罗人工智能实验室);Carnegie Mellon University(卡内基梅隆大学);Indraprastha Institute of Information Technology Delhi(德里英德拉普拉斯塔信息技术学院)
Department of Computer Science and Engineering, BRAC University(计算机科学与工程系,BRAC大学);Department of Computer Science and Engineering, Ahsanullah University of Science and Technology(计算机科学与工程系,Ahsanullah科技大学)