Trustworthy Text-to-Image Diffusion Models: A Timely and Focused Survey

2024-09-26

arXiv.org

AI Summary

The widespread adoption of text-to-image (T2I) diffusion models has exposed critical trustworthiness challenges—including robustness, fairness, privacy, and factual consistency—yet existing deep learning trustworthiness methods are ill-suited to their multimodal nature. Method: We propose the first structured taxonomy for T2I diffusion model trustworthiness, formally defining six core attributes and integrating novel analytical paradigms: falsification, enhancement, and verification & validation. We further introduce a multimodal-aware evaluation framework encompassing adversarial testing, interpretability analysis, bias detection, and factuality validation, accompanied by an open-source GitHub repository supporting empirical research. Contribution/Results: Our survey synthesizes over 12 mainstream benchmarks, four categories of analytical techniques, and cross-domain applications; identifies seven fundamental research gaps; and delivers both a methodological foundation and a practical roadmap for trustworthy AI-generated content.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Text-to-Image (T2I) Diffusion Models (DMs) have garnered widespread attention for their impressive advancements in image generation. However, their growing popularity has raised ethical and social concerns related to key non-functional properties of trustworthiness, such as robustness, fairness, security, privacy, factuality, and explainability, similar to those in traditional deep learning (DL) tasks. Conventional approaches for studying trustworthiness in DL tasks often fall short due to the unique characteristics of T2I DMs, e.g., the multi-modal nature. Given the challenge, recent efforts have been made to develop new methods for investigating trustworthiness in T2I DMs via various means, including falsification, enhancement, verification &validation and assessment. However, there is a notable lack of in-depth analysis concerning those non-functional properties and means. In this survey, we provide a timely and focused review of the literature on trustworthy T2I DMs, covering a concise-structured taxonomy from the perspectives of property, means, benchmarks and applications. Our review begins with an introduction to essential preliminaries of T2I DMs, and then we summarise key definitions/metrics specific to T2I tasks and analyses the means proposed in recent literature based on these definitions/metrics. Additionally, we review benchmarks and domain applications of T2I DMs. Finally, we highlight the gaps in current research, discuss the limitations of existing methods, and propose future research directions to advance the development of trustworthy T2I DMs. Furthermore, we keep up-to-date updates in this field to track the latest developments and maintain our GitHub repository at: https://github.com/wellzline/Trustworthy_T2I_DMs

Problem

• Address ethical concerns in text-to-image diffusion models
• Analyze trustworthiness properties like robustness and fairness
• Review methods for verifying and validating T2I model trustworthiness

Innovation

• Surveying trustworthiness in Text-to-Image Diffusion Models
• Analyzing non-functional properties via falsification and validation
• Proposing future directions for trustworthy T2I DMs