Automated Black-box Prompt Engineering for Personalized Text-to-Image Generation

TL;DR

This paper presents PRISM, an automated black-box prompt engineering algorithm that generates human-interpretable, transferable prompts for text-to-image models, reducing manual effort and improving cross-model applicability.

Contribution

Introduction of PRISM, a novel algorithm leveraging LLM in-context learning to automatically produce effective prompts for T2I models without white-box access.

Findings

PRISM generates accurate prompts for diverse objects and styles.

PRISM works across multiple T2I models including Stable Diffusion, DALL-E, and Midjourney.

PRISM produces human-interpretable prompts that are transferable.

Abstract

Prompt engineering is an effective but labor-intensive way to control text-to-image (T2I) generative models. Its time-intensive nature and complexity have spurred the development of algorithms for automated prompt generation. However, these methods often struggle with transferability across T2I models, require white-box access to the underlying model, or produce non-intuitive prompts. In this work, we introduce PRISM, an algorithm that automatically produces human-interpretable and transferable prompts that can effectively generate desired concepts given only black-box access to T2I models. Inspired by large language model (LLM) jailbreaking, PRISM leverages the in-context learning ability of LLMs to iteratively refine the candidate prompt distribution built upon the reference images. Our experiments demonstrate the versatility and effectiveness of PRISM in generating accurate prompts for objects, styles, and images across multiple T2I models, including Stable Diffusion, DALL-E, and Midjourney.