研究人员推出Dashboard2Code，一项旨在评估多模态大语言模型重建交互式仪表板能力的新任务。该任务超越了静态图表生成，要求模型与仪表板交互、整合反馈并生成复制目标仪表板的代码。为此，他们开发了DashboardMimic基准数据集，使用Plotly+Dash，包含180个不同难度级别和交互模式的仪表板-代码对。实验表明，包括强大的开源和闭源选项在内的当前模型在处理复杂仪表板时遇到困难，突显了显著的性能差距。 AI
影响 这项研究突显了当前多模态模型在复杂交互式数据可视化方面的局限性，指明了未来发展的方向。
排序理由 该集群描述了一篇介绍新颖任务和基准以评估多模态模型的研究论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2607.04727v1 Announce Type: cross Abstract: Automatic data visualization generation has advanced rapidly with multi-modal large language models, yet existing efforts largely focus on static charts and overlook the interactive dashboards commonly used for real-world data exp…
Automatic data visualization generation has advanced rapidly with multi-modal large language models, yet existing efforts largely focus on static charts and overlook the interactive dashboards commonly used for real-world data exploration. We introduce Dashboard2Code, a novel tas…
Automatic data visualization generation has advanced rapidly with multi-modal large language models, yet existing efforts largely focus on static charts and overlook the interactive dashboards commonly used for real-world data exploration. We introduce Dashboard2Code, a novel tas…