NuclearDiffusion: Text-to-Image Foundation Models for Learning Nuclear Energy Concepts

2026-08-01

AI Summary

This study addresses the limitations of general-purpose text-to-image models in specialized engineering domains such as nuclear energy, where a lack of domain-specific knowledge often leads to physically inaccurate or conceptually inconsistent outputs. The work presents the first systematic investigation into model adaptation for this domain, constructing a high-quality nuclear energy image dataset and applying supervised fine-tuning to open-source diffusion models—including Stable Diffusion XL, SD-v3.5-Medium, and Flux.1. Comprehensive evaluation, combining expert assessment and image similarity metrics, demonstrates that fine-tuning substantially enhances model performance in technical fidelity and physical consistency. Notably, under domain-specific prompts, the optimized SDXL outperforms commercial systems such as GPT-Image-2, Gemini, and Midjourney, underscoring the potential and advantages of adapting open-source models for specialized technical tasks.

Abstract

Generative artificial intelligence (AI) has transformed text-to-image synthesis, yet its ability to represent specialized engineering domains remains largely unexplored. As an exmaple in nuclear engineering, general-purpose foundation models frequently generate physically incorrect or conceptually inconsistent images because they lack domain-specific knowledge. This work presents one of the first systematic studies of domain adaptation for nuclear text-to-image generation through fine-tuning of open-source diffusion models. We curate a dataset of 1,000 captioned nuclear energy images spanning reactors, fuel cycles, radiation, and related concepts, and use it to fine-tune three state-of-the-art open-source models: Stable Diffusion XL (SDXL), SD-v3.5-Medium, and the flow-matching Flux.1 model. Their performance is evaluated using both quantitative image-similarity metrics and qualitative expert assessment against the corresponding zero-shot models. Fine-tuning substantially improves the fidelity of SDXL, provides only limited gains for SD-v3.5-Medium, and yields no measurable improvement for Flux.1, demonstrating that adaptation effectiveness depends strongly on the underlying generative architecture rather than model scale alone. We further compare the fine-tuned models against three leading commercial systems--GPT-Image-2, Gemini-3.1-Flash-Image, and Midjourney. Although GPT-Image-2 and Gemini generate convincing images for broad nuclear concepts, they frequently fail on specialized engineering prompts, where the fine-tuned open-source models produce more accurate and technically consistent outputs. These results establish domain-specific fine-tuning as a practical pathway for developing trustworthy generative AI tools for domain-specific applications.

Problem

• text-to-image generation
• nuclear engineering
• domain-specific knowledge
• generative AI
• foundation models

Innovation

• domain adaptation
• text-to-image generation
• nuclear engineering
• diffusion models
• fine-tuning