研究人员开发了两个新的图表到代码生成框架，旨在提高将视觉数据转换为可执行脚本的准确性和通用性。一种方法Chart2NCode引入了一个包含176,000个图表及其在Python、R和LaTeX中对齐脚本的数据集，以及一个名为CharLuMA的模型，该模型可以高效地适应不同的编程语言。另一个框架CharTide采用以数据为中心的方法，拥有一个200万个样本的数据集和一个查询驱动的RL框架，以增强视觉感知和代码逻辑，在与GPT-4o和GPT-5等模型相比时取得了有竞争力的结果。 AI
影响 图表到代码生成方面的进步可以简化数据可视化工作流程，并提高数据分析工具的可访问性。
排序理由 该集群包含两篇详细介绍图表到代码生成新框架和数据集的研究论文。
AI 生成摘要 · Google Gemini · 来自 5 个来源。 我们如何撰写摘要 →
arXiv:2601.14289v2 Announce Type: replace-cross Abstract: Understanding research papers remains challenging for foundation models due to specialized scientific discourse and complex figures and tables, yet existing benchmarks offer limited fine-grained evaluation at scale. To add…
arXiv:2604.24559v1 Announce Type: new Abstract: Chart-to-code generation converts a chart image into an executable plotting script, enabling faithful reproduction and editable visualizations. Existing methods are largely Python-centric, limiting practical use and overlooking a cr…
Chart-to-code generation converts a chart image into an executable plotting script, enabling faithful reproduction and editable visualizations. Existing methods are largely Python-centric, limiting practical use and overlooking a critical source of supervision: the same chart can…
arXiv:2604.22192v1 Announce Type: new Abstract: Chart-to-code generation demands strict visual precision and syntactic correctness from Vision-Language Models (VLMs). However, existing approaches are fundamentally constrained by data-centric limitations: despite the availability …
Chart-to-code generation demands strict visual precision and syntactic correctness from Vision-Language Models (VLMs). However, existing approaches are fundamentally constrained by data-centric limitations: despite the availability of growing chart-to-code datasets, simply scalin…