Learning to Sample Effective and Diverse Prompts for Text-to-Image Generation

TL;DR

This paper introduces PAG, a novel prompt adaptation method using GFlowNets for text-to-image generation, enabling diverse, high-quality prompts and overcoming limitations of reinforcement learning approaches.

Contribution

We propose PAG, a GFlowNet-based framework for prompt adaptation that improves diversity and effectiveness in text-to-image generation, addressing mode collapse and neural plasticity issues.

Findings

PAG generates more diverse prompts than RL-based methods.

PAG demonstrates robustness across different reward functions.

PAG transfers effectively to various text-to-image models.

Abstract

Recent advances in text-to-image diffusion models have achieved impressive image generation capabilities. However, it remains challenging to control the generation process with desired properties (e.g., aesthetic quality, user intention), which can be expressed as black-box reward functions. In this paper, we focus on prompt adaptation, which refines the original prompt into model-preferred prompts to generate desired images. While prior work uses reinforcement learning (RL) to optimize prompts, we observe that applying RL often results in generating similar postfixes and deterministic behaviors. To this end, we introduce \textbf{P}rompt \textbf{A}daptation with \textbf{G}FlowNets (\textbf{PAG}), a novel approach that frames prompt adaptation as a probabilistic inference problem. Our key insight is that leveraging Generative Flow Networks (GFlowNets) allows us to shift from reward maximization to sampling from an unnormalized density function, enabling both high-quality and diverse prompt generation. However, we identify that a naive application of GFlowNets suffers from mode collapse and uncovers a previously overlooked phenomenon: the progressive loss of neural plasticity in the model, which is compounded by inefficient credit assignment in sequential prompt generation. To address this critical challenge, we develop a systematic approach in PAG with flow reactivation, reward-prioritized sampling, and reward decomposition for prompt adaptation. Extensive experiments validate that PAG successfully learns to sample effective and diverse prompts for text-to-image generation. We also show that PAG exhibits strong robustness across various reward functions and transferability to different text-to-image models.