RULER：用于SVG生成的实例感知评分规则奖励

中文摘要 AI 辅助

从自然语言指令生成可缩放矢量图形（SVG）代码是一项开放式的任务，没有绝对的视觉真值，这使得评估和策略优化都缺乏可靠的信号。基于自然图像校准的标量指标（如CLIP、美学评分）在风格化的矢量内容上迁移效果不佳，并且将它们用作强化学习奖励会引发奖励黑客问题。我们通过基于评分规则的评分方法同时解决了这两个局限性。我们首先通过实验证明，使用多轴评分规则提示视觉语言评判者，其与人类判断的相关性远优于标量指标，无论是在样本间还是在指令内。基于这一发现，我们引入了RULER（用于强化学习的实例感知评分规则奖励），它将每条指令转换为一个包含六个项目的实例感知评分规则，涵盖语义、视觉和风格轴；评判者视觉语言模型逐项对渲染的生成结果进行评分，加权的满意度形成细粒度的奖励，并通过组相对策略优化进行优化。由于评分规则仅从文本中推导，RULER既不需要配对的SVG真值，也不需要人类偏好标签。在MMSVG-Illustration和MMSVG-Icon上，RULER将评分从0.432/0.395提升至0.693/0.683，超越了专门的SVG专家模型，并与规模大得多的DeepSeek-V3相当，消融实验表明评分规则设计是开放式SVG生成强化学习中的有效杠杆。项目页面可从此https URL访问。

英文摘要

Generating Scalable Vector Graphics (SVG) code from natural-language instructions is an open-ended task without absolute visual ground truth, leaving both evaluation and policy optimization without a faithful signal. Scalar metrics (CLIP, Aesthetic) calibrated on natural images transfer poorly to stylized vector content, and reusing them as RL rewards triggers reward hacking. We address both limitations with rubric-based scoring. We first establish empirically that prompting a vision-language judge with a multi-axis rubric correlates with human judgments far better than scalar metrics, both across samples and within instructions. Building on this finding, we introduce RULER (Instance-aware Rubric Rewards for Reinforcement Learning), which converts each instruction into an instance-aware rubric of six items spanning semantic, visual, and stylistic axes; a judge VLM scores rendered rollouts item-by-item, and the weighted satisfactions form a fine-grained reward optimized via Group Relative Policy Optimization. Because the rubric is derived from text alone, RULER requires neither paired SVG ground truth nor human preference labels. On MMSVG-Illustration and MMSVG-Icon, RULER lifts the rubric score from 0.432/0.395 to 0.693/0.683, surpassing dedicated SVG specialists and matching the substantially larger DeepSeek-V3, with ablations identifying rubric design as the active lever for RL on open-ended SVG generation. The project page is available at https://hangyuran.github.io/RULER/.

发表机构

• Ant Group（蚂蚁集团）
• The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））
• University of Oxford（牛津大学）

机构由 AI 辅助整理，请以论文原文为准。