RULER:用实例化评分细则奖励提升 SVG 生成
RULER: Instance-aware Rubric Rewards for SVG Generation
一篇教 RL 生成 SVG 的论文：把每条指令拆成 6 项评分细则当奖励，不需要标注数据，效果追平 DeepSeek-V3，做生成任务奖励设计的可以看看。
从自然语言指令生成 SVG 代码缺乏绝对视觉基准，CLIP、Aesthetic 等标量指标迁移到矢量内容效果差，直接用作 RL 奖励还容易触发 reward hacking。论文提出 RULER，将每条指令转成涵盖语义、视觉、风格三个维度的 6 项实例化评分细则，由 judge VLM 逐项打分，通过 GRPO 优化加权满意度作为奖励。由于评分细则仅从文本推导，RULER 不需要 SVG 真值配对或人类偏好标注。在 MMSVG-Illustration 和 MMSVG-Icon 上，rubric 分数从 0.432/0.395 提升到 0.693/0.683，超过专门的 SVG 模型，并追平规模大得多的 DeepSeek-V3。
RULER: Instance-aware Rubric Rewards for SVG Generation
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/.