SciForma: 結構忠實的科學圖表生成

摘要

結構保真度對科學方法論圖表至關重要。為傳達研究邏輯，這些圖表須忠實呈現元件、方向關係及文字註解。由於單一錯誤（例如箭頭反向或方程式難以辨識）可能使整個圖形失效，結構保真本質上具有合取性：在某個面向上的正確性無法補償另一面向的失誤。現有開源模型未能滿足此項標準。監督式微調（SFT）可學習合理的佈局，但無法可靠確保結構正確性；而基於標量獎勵的後訓練則模糊了哪個結構面向發生錯誤。為解決此問題，我們提出 SciForma，一個用於科學方法論圖表結構保真生成的框架。具體而言，SciForma 將圖表品質分解為三個結構軸向：元件、箭頭與文字，並以結構清單為引導。在此基礎上，我們策劃了 SciFormaData-700K 用於結構化訓練，以及 SciFormaBench-2K 用於邏輯驗證評估。為填補 SFT 留下的差距，我們開發了多維合取偏好最佳化（M-DPO），該方法強制所有軸向同時正確，並在後訓練中自適應地將梯度導向最不足的維度。相同的結構清單還能於推論時進行疊代編輯，以修正殘留錯誤。此組合使 SciForma-9B 在 SciFormaBench-2K 與 AIBench 上均超越所有開源基準及 GPT-Image-1.5，將開放科學圖表生成提升至接近專有等級的結構保真度。我們的程式碼與數據將公布於：https://github.com/microsoft/SciForma。

English

Structural fidelity is essential to scientific methodology diagrams. To communicate research logic, these diagrams must faithfully render components, directional relations, and textual annotations. Since a single error, such as a reversed arrow or an unreadable equation, can invalidate the entire figure, structural fidelity is inherently conjunctive: correctness on one axis cannot compensate for failure on another. Current open-source models fail to satisfy this criterion. Supervised fine-tuning (SFT) learns plausible layouts but cannot reliably ensure structural correctness, while scalar reward-based post-training obscures which structural dimension has failed. To address this, we introduce SciForma, a framework for the structure faithful generation of scientific methodology diagrams. Specifically, SciForma decomposes diagram quality into three structural axes: Component, Arrow, and Text, guided by a structural inventory. Built on this foundation, we curate SciFormaData-700K for structured training and SciFormaBench-2K for logic-verified evaluation. To close the gap left by SFT, we develop Multi-Dimensional Conjunctive Preference Optimization (M-DPO), which enforces simultaneous correctness across all axes and adaptively routes gradients to the most deficient dimension in post-training. The same structural inventory also enables iterative editing at inference time to correct residual errors. This combination allows SciForma-9B to exceed all open-source baselines and GPT-Image-1.5 on both SciFormaBench-2K and AIBench, bringing open scientific diagram generation close to proprietary-level structural fidelity. Our code and data will be available at: https://github.com/microsoft/SciForma.