Reason-SVG: Hybrid Reward RL for Aha-Moments in Vector Graphics Generation
This work addresses a domain-specific problem for AI-assisted design tools by enhancing LLM reasoning for SVG generation, though it appears incremental as it builds on existing methods like SFT and RL with novel adaptations.
The paper tackles the challenge of generating high-quality Scalable Vector Graphics (SVGs) with Large Language Models (LLMs) by introducing Reason-SVG, a framework that combines a 'Drawing-with-Thought' paradigm with a two-stage training strategy using supervised fine-tuning and reinforcement learning with a hybrid reward function, resulting in significant improvements in generating accurate and visually compelling SVGs.
Generating high-quality Scalable Vector Graphics (SVGs) is challenging for Large Language Models (LLMs), as it requires advanced reasoning for structural validity, semantic faithfulness, and visual coherence -- capabilities in which current LLMs often fall short. In this work, we introduce Reason-SVG, a novel framework designed to enhance LLM reasoning for SVG generation. Reason-SVG pioneers the "Drawing-with-Thought" (DwT) paradigm, in which models generate both SVG code and explicit design rationales, mimicking the human creative process. Reason-SVG adopts a two-stage training strategy: First, Supervised Fine-Tuning (SFT) trains the LLM on the DwT paradigm to activate foundational reasoning abilities. Second, Reinforcement Learning (RL), utilizing Group Relative Policy Optimization (GRPO), empowers the model to generate both DwT and SVGs rationales through refined, reward-driven reasoning. To facilitate reasoning-driven SVG generation, we design a Hybrid Reward function that evaluates the presence and utility of DwT reasoning, along with structural validity, semantic alignment, and visual quality. We also introduce the SVGX-DwT-10k dataset, a high-quality corpus of 10,000 SVG-DwT pairs, where each SVG code is generated based on explicit DwT reasoning. By integrating DwT, SFT, and Hybrid Reward-guided RL, Reason-SVG significantly improves LLM performance in generating accurate and visually compelling SVGs, potentially fostering "Aha moments" in design.