AI 中文总结
本研究提出形式化感知奖励循环，通过微调视觉-语言模型结合Group Relative Policy Optimisation生成PlantUML，提升了转换质量，为建模评估提供了新方向。
AI 中文摘要
手写UML草图在早期软件设计中很常见，但将其转化为结构化、可分析的建模制品仍需手动重建。视觉-语言模型可从图表图像生成PlantUML，但基于提示词的使用将其视为图像到文本生成，而非结构化模型生成。我们研究形式化感知奖励：源自可分析模型表示而非表面文本的反馈信号。在一个工作示例中，我们通过监督微调后接Group Relative Policy Optimisation，适配了一个视觉-语言模型用于手写UML到PlantUML的生成。将生成的PlantUML与目标表示进行比较，类图使用XMI，活动图使用控制流图。初步结果显示，适配后的模型相比未微调的开源模型和一个专有基线，在可编译性和转换质量上有所提升，同时在类图上与更强的专有基线表现相当。奖励引导阶段在当前保留集上的额外收益仍不明确。错误分析和指标有效性结果表明，建模可接受性仅被部分捕捉，这推动了将模型分析与人类判断相结合的奖励和评估方法。
英文摘要
Handwritten UML sketches are common in early software design, but turning them into structured, analysable modelling artefacts still requires manual reconstruction. Vision-language models can generate PlantUML from diagram images, but prompt-based use treats this as image-to-text generation rather than structured model generation. We investigate formalism-aware rewards: feedback signals derived from analysable model representations rather than surface text. In a worked example, we adapt a vision-language model for handwritten UML-to-PlantUML generation using super-vised fine-tuning followed by Group Relative Policy Optimisation. Generated PlantUML is compared against target representations, using XMI for class diagrams and control-flow graphs for activity diagrams. Emerging results show that the adapted model improves compilability and conversion quality over the untuned open model and one proprietary baseline, while remaining competitive with a stronger proprietary baseline on class diagrams. The added benefit of the reward-guided stage remains open on the current held-out set. Error analysis and metric-validity results show that modelling acceptability is only partially captured, motivating rewards and evaluations that combine model analysis with human judgement.
CommentsAccepted for publication in the Proceedings of the ACM/IEEE 29th International Conference on Model Driven Engineering Languages and Systems (MODELS 2026). This is the accepted author manuscript
