REChart: Reasoning-Efficient Chart Editing with Large Reasoning Models

📄 arXiv: 2608.17414v1 📥 PDF

作者: Yuanbang Liu, Chenxi Ruan, Yihan Hou, Qiong Luo, Wei Zeng

分类: cs.CV, cs.PL

发布日期: 2026-08-18


💡 一句话要点

提出REChart以解决图表编辑中的推理效率问题

🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture) 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 图表编辑 推理效率 多模态任务 强化学习 过程级监督

📋 核心要点

  1. 现有方法在图表编辑中面临推理效率低下和过度思考的问题,导致性能下降。
  2. REChart通过两阶段训练框架,提供过程级监督,优化推理步骤,提高编辑的准确性和效率。
  3. 在ChartEdit和ChartMIMIC基准测试中,REChart在开放源代码模型中实现了最先进的图表编辑性能,同时将平均推理令牌使用量减少了79.0%。

📝 摘要(中文)

图表编辑需要根据参考图表图像和编辑指令推断和修改可视化代码,这对细粒度视觉推理、指令遵循和可执行代码合成能力提出了挑战。大型推理模型(LRMs)结合扩展的链式思维(CoT)推理适合处理此类复杂的多模态任务。然而,初步研究表明推理长度与图表编辑性能之间存在“倒U型”关系:过度推理往往导致“过度思考”,模型可能偏离真实视觉细节或陷入冗余推理循环。为了解决这一问题,本文提出了REChart,一个两阶段训练框架,通过对中间推理步骤的过程级监督,提高编辑的准确性和推理效率。

🔬 方法详解

问题定义:本文旨在解决图表编辑中推理效率低和过度思考的问题。现有方法在处理复杂的多模态任务时,往往导致模型偏离真实细节或陷入冗余推理循环。

核心思路:REChart的核心思路是通过两阶段的训练框架,提供对中间推理步骤的监督,从而提高推理效率和编辑准确性。通过角色专门化的Reason-Score-Refine工作流,逐步优化图表代码。

技术框架:REChart的整体架构包括两个主要阶段:首先,从一个大型图像-指令-代码池中合成20万条高质量推理轨迹进行监督微调;其次,通过强化学习优化模型,使用忠实度和效率两个互补奖励进行训练。

关键创新:REChart的创新点在于引入了过程级监督和强化学习相结合的训练方法,显著提高了推理效率,避免了过度思考现象。与现有方法相比,REChart在推理长度和性能之间找到了更好的平衡。

关键设计:在设计中,使用了特定的损失函数来评估代码的正确性、视觉保真度和结构一致性,同时为每次推理分配随机的思考预算,以优化推理过程。

🖼️ 关键图片

fig_0
fig_1

📊 实验亮点

REChart在ChartEdit和ChartMIMIC基准测试中表现出色,成为开放源代码模型中图表编辑性能的领先者。与基础模型相比,在最大思考预算为16,384个令牌的情况下,平均推理令牌使用量减少了79.0%,有效缓解了过度思考问题。

🎯 应用场景

该研究的潜在应用领域包括数据可视化、自动化报告生成和人机交互界面设计等。REChart的高效推理能力和准确的图表编辑功能可以显著提升数据分析和展示的效率,未来可能在商业智能和教育领域产生深远影响。

📄 摘要(原文)

Chart editing requires inferring and modifying visualization code from a reference chart image based on an editing instruction, challenging fine-grained visual reasoning, instruction following, and executable code synthesis capabilities of MLLMs. Large reasoning models (LRMs) with extended Chain-of-Thought (CoT) reasoning are suitable for tackling such complex multimodal tasks. However, our preliminary study reveals an inverted-U'' relationship between reasoning length and chart-editing performance: Excessive reasoning often leads tooverthinking,'' where models drift toward hallucinated visual details or get stuck in redundant reasoning loops. To address the gap, we introduce REChart, a two-stage training framework that provides process-level supervision over intermediate reasoning steps, improving both editing fidelity and reasoning efficiency. First, we synthesize 200k high-quality reasoning trajectories for supervised fine-tuning from a large image-instruction-code pool, using a role-specialized agentic Reason-Score-Refine workflow that iteratively refine the chart code toward higher quality. Second, we optimize the model via reinforcement learning with two complementary rewards: a \emph{fidelity} reward evaluating code correctness, visual fidelity, and structural consistency, and an \emph{efficiency} reward that assigns each rollout a random thinking budget, truncates the reasoning process, and credits the final reasoning segment according to its contribution to the output. On the ChartEdit and ChartMIMIC benchmarks, our model achieves state-of-the-art chart-editing performance among open-source models of comparable scale, while mitigating overthinking and reducing average reasoning token usage by 79.0\% under a maximum thinking budget of 16,384 tokens compared with the base model.