AI Wizards at EXIST 2026: Hierarchical Soft-Label Learning for Multimodal Sexism Identification in Memes
- 类型:arxiv
- 标识:2607.04410
- 链接:https://arxiv.org/abs/2607.04410
- 主分类:multimodal
- 形态:method
- 被引:0
- 被引来源:Semantic Scholar + OpenAlex
- S2被引:0
- OpenAlex被引:0
- 影响力被引:0
- TLDR:This work maps fixed Gemini Embedding 2 vision-language representations through a lightweight Gated MLP trained with KL divergence and homoscedastic uncertainty weighting, and presents the AI Wizards submission to EXIST 2026 for multimodal sexism identification in memes.
- OpenAlex ID:W7167635211
- OpenAlex DOI:10.48550/arxiv.2607.04410
- DOI:10.48550/arxiv.2607.04410
- DOI来源:OpenAlex
- 开放获取:green
- 开放获取链接:https://doi.org/10.48550/arxiv.2607.04410
- OpenAlex更新:2026-07-19
- 待LLM分类:否
- 标题中文:AI Wizards 参加 EXIST 2026:用于迷因中多模态性别歧视识别的分层软标签学习
- TLDR中文:该工作通过轻量级 Gated MLP 将固定的 Gemini Embedding 2 视觉-语言表征映射到目标空间,使用 KL 散度与同方差不确定性加权进行训练,并展示了 AI Wizards 参加 EXIST 2026 多模态迷因性别歧视识别任务的提交方案。
- 来源文件:
- /inbox/tom/_candidates/2026-07-07-agent-rag-longcontext-candidates.json
- [S2 enrich]
- [OpenAlex backfill]