Reliable Reasoning in SVG-LLMs via Multi-Task Multi-Reward Reinforcement Learning

2026-03-17

AI Summary

Existing methods for SVG generation suffer from limited generalization, redundant code output, and a lack of explicit reasoning mechanisms. To address these limitations, this work proposes CTRL-S, a novel framework that introduces chain-of-thought (CoT) reasoning into SVG generation for the first time, explicitly modeling structured decision pathways during the generation process. We construct SVG-Sophia, a high-quality multitask dataset, and employ Group Relative Policy Optimization (GRPO) to jointly optimize multiple reward signals—including DINO visual features, image-text similarity, format compliance, and code efficiency. Experimental results demonstrate that CTRL-S significantly outperforms current approaches in task success rate, SVG code quality, and visual fidelity.

Technology Category

• Computer Vision: Visual Reasoning & Symbolic Representations
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

With the rapid advancement of vision-language models, an increasing number of studies have explored their potential for SVG generation tasks. Although existing approaches improve performance by constructing large-scale SVG datasets and introducing SVG-specific tokens, they still suffer from limited generalization, redundant paths in code outputs, and a lack of explicit reasoning. In this work, we present CTRL-S (Chain-of-Thought Reinforcement Learning for SVG), a unified framework that introduces a chain-of-thought mechanism to explicitly expose the model's reasoning process during SVG generation. To support this structured reasoning, we construct SVG-Sophia, a high-quality dataset containing 145K samples across SVG code refinement, Text-to-SVG, and Image-to-SVG tasks. By training the model to generate group-level structured SVG code, CTRL-S significantly improves structural coherence and visual fidelity. Furthermore, we adopt the GRPO algorithm and design a multi-reward optimization framework, incorporating DINO, image-text similarity, format, and code efficiency rewards. Through joint multi-reward optimization and multi-task training, our approach systematically enhances overall generation capabilities. Extensive experiments show that CTRL-S outperforms existing methods, achieving higher task success rates, superior SVG code quality, and exceptional visual fidelity.

Problem

• SVG generation
• limited generalization
• redundant paths
• explicit reasoning
• vision-language models

Innovation

• Chain-of-Thought Reasoning
• Multi-Reward Reinforcement Learning
• Structured SVG Generation
• Multi-Task Training
• GRPO Algorithm