From Sparse to Dense: Multi-View GRPO for Flow Models via Augmented Condition Space

TL;DR

This paper introduces Multi-View GRPO, a novel method that enhances preference alignment in text-to-image flow models by augmenting the condition space to enable richer inter-sample relationship exploration, leading to improved performance.

Contribution

The paper proposes MV-GRPO, which uses a Condition Enhancer to generate diverse captions, enabling multi-view reward estimation without additional sample generation, thus improving alignment.

Findings

MV-GRPO outperforms state-of-the-art methods in alignment tasks.

Enhanced relationship exploration leads to better semantic attribute capture.

The approach improves performance without costly sample regeneration.

Abstract

Group Relative Policy Optimization (GRPO) has emerged as a powerful framework for preference alignment in text-to-image (T2I) flow models. However, we observe that the standard paradigm where evaluating a group of generated samples against a single condition suffers from insufficient exploration of inter-sample relationships, constraining both alignment efficacy and performance ceilings. To address this sparse single-view evaluation scheme, we propose Multi-View GRPO (MV-GRPO), a novel approach that enhances relationship exploration by augmenting the condition space to create a dense multi-view reward mapping. Specifically, for a group of samples generated from one prompt, MV-GRPO leverages a flexible Condition Enhancer to generate semantically adjacent yet diverse captions. These captions enable multi-view advantage re-estimation, capturing diverse semantic attributes and providing richer optimization signals. By deriving the probability distribution of the original samples conditioned on these new captions, we can incorporate them into the training process without costly sample regeneration. Extensive experiments demonstrate that MV-GRPO achieves superior alignment performance over state-of-the-art methods.