Text2Arch: A Dataset for Generating Scientific Architecture Diagrams from Natural Language Descriptions

2026-04-16 · Source: Computation and Language · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Advanced, quick

Summary

A new dataset named Text2Arch has been introduced to facilitate the automatic generation of scientific architecture diagrams from natural language descriptions. This dataset includes scientific architecture images, their corresponding textual descriptions, and associated DOT code representations. The goal is to address the inefficiency and ambiguity of communicating complex system designs through text alone, with potential applications in enterprise architecture visualization, AI-driven software design, and educational content creation. Researchers fine-tuned a suite of small language models and utilized in-context learning with GPT-4o, demonstrating that models trained on Text2Arch significantly outperform existing baselines like DiagramAgent and achieve performance comparable to GPT-4o's in-context learning generations. The code, data, and models are publicly available.

Key takeaway

For research scientists developing AI systems for design and visualization, Text2Arch offers a critical resource for training models to generate architecture diagrams from text. You should explore fine-tuning small language models with this dataset to achieve performance comparable to large models like GPT-4o, potentially reducing inference costs and improving accessibility for specific applications.

Key insights

Text2Arch dataset enables language models to generate scientific architecture diagrams from text.

Principles

• Text-to-diagram generation improves clarity.
• Large-scale datasets are crucial for model training.

Method

The method involves semantic understanding of input text to generate intermediate DOT code, which is then processed to create high-fidelity architecture diagrams.

In practice

• Use Text2Arch for AI-driven software design.
• Apply Text2Arch in educational content creation.

Topics

• Text2Arch Dataset
• Architecture Diagram Generation
• Language Models
• DOT Code
• Semantic Understanding

Best for: Research Scientist, AI Scientist, Machine Learning Engineer, NLP Engineer

Related on AIssential

• Text2CAD-Bench: A Benchmark for LLM-based Text-to-Parametric CAD Generation — Text2CAD-Bench systematically evaluates LLM-based text-to-CAD generation across diverse geometric complexities and real-world applications.
• Beyond One Output: Visualizing and Comparing Distributions of Language Model Generations — Visualizing language model output distributions as interactive graphs reveals hidden structures and improves diversity assessment.
• PubMed-Ophtha: An open resource for training ophthalmology vision-language models on scientific literature — PubMed-Ophtha provides a high-quality, large-scale ophthalmology image-text dataset for vision-language model development.
• Building LLM From Scratch: Understanding How Large Language Models Work — Large Language Models fundamentally predict the next token by processing numerical representations of text through attention mechanisms.
• Faithful, Enriched, and Precise: Benchmarking Natural-Science Illustration Generation by T2I models — FEPBench offers a fine-grained, atom-set-based benchmark for T2I models generating scientific illustrations, revealing current limitations in text, reasoning, and precision.
• ChartNet: A Million-Scale, High-Quality Multimodal Dataset for Robust Chart Understanding — Code-guided synthetic data generation significantly improves VLM chart understanding and reasoning capabilities.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Computation and Language.