研究人员正在开发新的方法来评估和生成人工智能幽默，解决什么使事物有趣的主观性质。一种方法涉及使用偏好建模的“生成多个，选择最佳”策略，该策略在 SemEval-2026 幽默生成任务中排名很高。另一项研究分析了 YouTube Shorts 上的黑暗幽默和常规幽默，发现黑暗幽默会引起更多混合甚至有毒的观众反应。此外，还创建了一个名为 v-HUB 的新基准来评估多模态 LLM 理解视频中幽默的能力，并强调包含音频可以提高理解能力。 AI
影响 新的基准和方法正在出现，以评估人工智能对幽默的细微理解和生成能力，从而推动多模态和约束式人工智能能力的边界。
排序理由 多篇研究论文介绍了使用人工智能评估和生成幽默的新基准、数据集和方法。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2606.00022v1 Announce Type: cross Abstract: Humor generation remains difficult not only because producing fluent, novel jokes is hard, but because "funny" is audience-dependent and supervision is noisy -- preferences vary with audience, context, and culture, and annotator a…
arXiv:2606.00046v1 Announce Type: cross Abstract: Video platforms such as YouTube have reshaped how users engage with entertainment and information, emphasizing brief, highly engaging content such as Shorts. Within this ecosystem, certain content occupies a gray area where it rem…
arXiv:2509.25773v3 Announce Type: replace-cross Abstract: AI models capable of comprehending humor hold real-world promise -- for example, enhancing engagement in human-machine interactions. To gauge and diagnose the capacity of multimodal large language models (MLLMs) for humor …
arXiv:2604.19786v2 Announce Type: replace Abstract: Humor remains difficult to evaluate in large language models (LLMs) because what makes a response funny is subjective, comparative, and shaped by interacting comedic mechanisms rather than a single scalar property. Existing humo…