SciFig: Towards Automating Scientific Figure Generation

2026-01-07

arXiv.org

AI Summary

This study addresses the labor-intensive and expertise-dependent process of generating high-quality flowcharts from scientific papers. To overcome this challenge, the authors propose the first end-to-end AI agent system capable of automatically constructing publication-ready diagrams directly from paper text. The approach innovatively integrates hierarchical module layout generation, functional clustering, and connection modeling, augmented by an iterative chain-of-thought (CoT) visual reasoning mechanism and a rule-based automated evaluation framework. Evaluated on both general and paper-specific datasets, the system achieves overall quality scores of 70.1% and 66.2%, respectively, demonstrating strong performance in visual clarity, structural organization, and scientific accuracy.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Knowledge Representation and Reasoning: Automated Reasoning and Theorem Proving

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Creating high-quality figures and visualizations for scientific papers is a time-consuming task that requires both deep domain knowledge and professional design skills. Despite over 2.5 million scientific papers published annually, the figure generation process remains largely manual. We introduce $\textbf{SciFig}$, an end-to-end AI agent system that generates publication-ready pipeline figures directly from research paper texts. SciFig uses a hierarchical layout generation strategy, which parses research descriptions to identify component relationships, groups related elements into functional modules, and generates inter-module connections to establish visual organization. Furthermore, an iterative chain-of-thought (CoT) feedback mechanism progressively improves layouts through multiple rounds of visual analysis and reasoning. We introduce a rubric-based evaluation framework that analyzes 2,219 real scientific figures to extract evaluation rubrics and automatically generates comprehensive evaluation criteria. SciFig demonstrates remarkable performance: achieving 70.1$\%$ overall quality on dataset-level evaluation and 66.2$\%$ on paper-specific evaluation, and consistently high scores across metrics such as visual clarity, structural organization, and scientific accuracy. SciFig figure generation pipeline and our evaluation benchmark will be open-sourced.

Problem

• scientific figure generation
• visualization
• automated design
• publication-ready figures
• manual process

Innovation

• scientific figure generation
• hierarchical layout generation
• iterative chain-of-thought feedback
• rubric-based evaluation
• AI agent system