EAM speeds up adjoint matching for diffusion model reward fine-tuning by switching to linear base drift, allowing deterministic few-step solvers and closed-form adjoints with up to 4x faster convergence on text-to-image benchmarks.
Pref-GRPO: Pairwise Preference Reward-based GRPO for Stable Text-to-Image Reinforcement Learning
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
abstract
Recent advancements highlight the importance of GRPO-based reinforcement learning methods and benchmarking in enhancing text-to-image (T2I) generation. However, current methods using pointwise reward models (RM) for scoring generated images are susceptible to reward hacking. We reveal that this happens when minimal score differences between images are amplified after normalization, creating illusory advantages that drive the model to over-optimize for trivial gains, ultimately destabilizing the image generation process. To address this, we propose Pref-GRPO, a pairwise preference reward-based GRPO method that shifts the optimization objective from score maximization to preference fitting, ensuring more stable training. In Pref-GRPO, images are pairwise compared within each group using preference RM, and the win rate is used as the reward signal. Extensive experiments demonstrate that PREF-GRPO differentiates subtle image quality differences, providing more stable advantages and mitigating reward hacking. Additionally, existing T2I benchmarks are limited by coarse evaluation criteria, hindering comprehensive model assessment. To solve this, we introduce UniGenBench, a unified T2I benchmark comprising 600 prompts across 5 main themes and 20 subthemes. It evaluates semantic consistency through 10 primary and 27 sub-criteria, leveraging MLLM for benchmark construction and evaluation. Our benchmarks uncover the strengths and weaknesses of both open and closed-source T2I models and validate the effectiveness of Pref-GRPO.
citation-role summary
background 2
citation-polarity summary
years
2026 9verdicts
UNVERDICTED 9roles
background 2representative citing papers
TMPO uses Softmax Trajectory Balance to match policy probabilities over multiple trajectories to a Boltzmann reward distribution, improving diversity by 9.1% in diffusion alignment tasks.
HP-Edit introduces a post-training framework and RealPref-50K dataset that uses a VLM-based HP-Scorer to align diffusion image editing models with human preferences, improving outputs on Qwen-Image-Edit-2509.
TabGRAA enables self-improving tabular language models through iterative group-relative advantage alignment using modular automated quality signals like distinguishability classifiers.
LeapAlign fine-tunes flow matching models by constructing two consecutive leaps that skip multiple ODE steps with randomized timesteps and consistency weighting, enabling stable updates at any generation step.
Super-Linear Advantage Shaping (SLAS) introduces a non-linear geometric policy update for RL post-training of text-to-image models that reshapes the local policy space via advantage-dependent Fisher-Rao weighting to reduce reward hacking and improve performance over GRPO baselines.
Edit-R1 trains a CoT-based reasoning reward model with GCPO and uses it to boost image editing performance over VLMs and models like FLUX.1-kontext via GRPO.
RATS lets few-step visual generators surpass multi-step teachers by shaping trajectories with reward-based adaptive guidance instead of strict imitation.
The paper introduces the Proxy Compression Hypothesis as a unifying framework explaining reward hacking in RLHF as an emergent result of compressing high-dimensional human objectives into proxy reward signals under optimization pressure.
citing papers explorer
-
Efficient Adjoint Matching for Fine-tuning Diffusion Models
EAM speeds up adjoint matching for diffusion model reward fine-tuning by switching to linear base drift, allowing deterministic few-step solvers and closed-form adjoints with up to 4x faster convergence on text-to-image benchmarks.
-
TMPO: Trajectory Matching Policy Optimization for Diverse and Efficient Diffusion Alignment
TMPO uses Softmax Trajectory Balance to match policy probabilities over multiple trajectories to a Boltzmann reward distribution, improving diversity by 9.1% in diffusion alignment tasks.
-
HP-Edit: A Human-Preference Post-Training Framework for Image Editing
HP-Edit introduces a post-training framework and RealPref-50K dataset that uses a VLM-based HP-Scorer to align diffusion image editing models with human preferences, improving outputs on Qwen-Image-Edit-2509.
-
Self-Improving Tabular Language Models via Iterative Group Alignment
TabGRAA enables self-improving tabular language models through iterative group-relative advantage alignment using modular automated quality signals like distinguishability classifiers.
-
LeapAlign: Post-Training Flow Matching Models at Any Generation Step by Building Two-Step Trajectories
LeapAlign fine-tunes flow matching models by constructing two consecutive leaps that skip multiple ODE steps with randomized timesteps and consistency weighting, enabling stable updates at any generation step.
-
Power Reinforcement Post-Training of Text-to-Image Models with Super-Linear Advantage Shaping
Super-Linear Advantage Shaping (SLAS) introduces a non-linear geometric policy update for RL post-training of text-to-image models that reshapes the local policy space via advantage-dependent Fisher-Rao weighting to reduce reward hacking and improve performance over GRPO baselines.
-
Leveraging Verifier-Based Reinforcement Learning in Image Editing
Edit-R1 trains a CoT-based reasoning reward model with GCPO and uses it to boost image editing performance over VLMs and models like FLUX.1-kontext via GRPO.
-
Reward-Aware Trajectory Shaping for Few-step Visual Generation
RATS lets few-step visual generators surpass multi-step teachers by shaping trajectories with reward-based adaptive guidance instead of strict imitation.
-
Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges
The paper introduces the Proxy Compression Hypothesis as a unifying framework explaining reward hacking in RLHF as an emergent result of compressing high-dimensional human objectives into proxy reward signals under optimization pressure.