Reward Sharpness-Aware Fine-Tuning for Diffusion Models

• URL: http://arxiv.org/abs/2603.21175v1
• Date: Sun, 22 Mar 2026 11:32:50 GMT
• Title: Reward Sharpness-Aware Fine-Tuning for Diffusion Models
• Abstract summary: Reinforcement learning from human feedback has proven effective in aligning large language models with human preferences.<n>RDRL remains susceptible to reward hacking, where the reward score increases without corresponding improvements in perceptual quality.<n>We introduce methods that exploit gradients from a robustified reward model without requiring its retraining.
• Score: 10.026496861838448
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reinforcement learning from human feedback (RLHF) has proven effective in aligning large language models with human preferences, inspiring the development of reward-centric diffusion reinforcement learning (RDRL) to achieve similar alignment and controllability. While diffusion models can generate high-quality outputs, RDRL remains susceptible to reward hacking, where the reward score increases without corresponding improvements in perceptual quality. We demonstrate that this vulnerability arises from the non-robustness of reward model gradients, particularly when the reward landscape with respect to the input image is sharp. To mitigate this issue, we introduce methods that exploit gradients from a robustified reward model without requiring its retraining. Specifically, we employ gradients from a flattened reward model, obtained through parameter perturbations of the diffusion model and perturbations of its generated samples. Empirically, each method independently alleviates reward hacking and improves robustness, while their joint use amplifies these benefits. Our resulting framework, RSA-FT (Reward Sharpness-Aware Fine-Tuning), is simple, broadly compatible, and consistently enhances the reliability of RDRL.

Related papers

• Mitigating Reward Hacking in RLHF via Bayesian Non-negative Reward Modeling [49.41422138354821]
  We propose a principled reward modeling framework that integrates non-negative factor analysis into the Bradley-Terry preference model.<n>BNRM represents rewards through a sparse, non-negative latent factor generative process.<n>We show that BNRM substantially mitigates reward over-optimization, improves robustness under distribution shifts, and yields more interpretable reward decompositions than strong baselines.
  arXiv  Detail & Related papers  (2026-02-11T08:14:11Z)
• FinPercep-RM: A Fine-grained Reward Model and Co-evolutionary Curriculum for RL-based Real-world Super-Resolution [87.57784204422218]
  Reinforcement Learning with Human Feedback has proven effective in image generation field guided by reward models to align human preferences.<n>We propose a Fine-grained Perceptual Reward Model (FinPercep-RM) based on ancoder-Decoder architecture.<n>While providing a global quality score, it also generates a Perceptual Degradation Map that spatially localizes and quantifies local defects.
  arXiv  Detail & Related papers  (2025-12-27T16:55:21Z)
• Data-regularized Reinforcement Learning for Diffusion Models at Scale [99.01056178660538]
  We introduce Data-regularized Diffusion Reinforcement Learning ( DDRL), a novel framework that uses the forward KL divergence to anchor the policy to an off-policy data distribution.<n>With over a million GPU hours of experiments and ten thousand double-blind evaluations, we demonstrate that DDRL significantly improves rewards while alleviating the reward hacking seen in RLs.
  arXiv  Detail & Related papers  (2025-12-03T23:45:07Z)
• Enhancing Diffusion-based Restoration Models via Difficulty-Adaptive Reinforcement Learning with IQA Reward [93.04811239892852]
  Reinforcement Learning (RL) has recently been incorporated into diffusion models.<n>In this paper, we investigate how to effectively integrate RL into diffusion-based restoration models.
  arXiv  Detail & Related papers  (2025-11-03T14:57:57Z)
• Reward Models Can Improve Themselves: Reward-Guided Adversarial Failure Mode Discovery for Robust Reward Modeling [27.11560841914813]
  We introduce REFORM, a self-improving reward modeling framework that enhances robustness by using the reward model itself to guide the generation of falsely scored responses.<n>We evaluate REFORM on two widely used preference datasets Anthropic Helpful Harmless (HH) and PKU Beavertails.
  arXiv  Detail & Related papers  (2025-07-08T21:56:33Z)
• VARD: Efficient and Dense Fine-Tuning for Diffusion Models with Value-based RL [28.95582264086289]
  VAlue-based Reinforced Diffusion (VARD) is a novel approach that first learns a value function predicting expection of rewards from intermediate states.<n>Our method maintains proximity to the pretrained model while enabling effective and stable training via backpropagation.
  arXiv  Detail & Related papers  (2025-05-21T17:44:37Z)
• ROCM: RLHF on consistency models [8.905375742101707]
  We propose a reward optimization framework for applying RLHF to consistency models.<n>We investigate various $f$-divergences as regularization strategies, striking a balance between reward and model consistency.
  arXiv  Detail & Related papers  (2025-03-08T11:19:48Z)
• Fine-Tuning of Continuous-Time Diffusion Models as Entropy-Regularized Control [54.132297393662654]
  Diffusion models excel at capturing complex data distributions, such as those of natural images and proteins. While diffusion models are trained to represent the distribution in the training dataset, we often are more concerned with other properties, such as the aesthetic quality of the generated images. We present theoretical and empirical evidence that demonstrates our framework is capable of efficiently generating diverse samples with high genuine rewards.
  arXiv  Detail & Related papers  (2024-02-23T08:54:42Z)
• Low-Light Image Enhancement with Wavelet-based Diffusion Models [50.632343822790006]
  Diffusion models have achieved promising results in image restoration tasks, yet suffer from time-consuming, excessive computational resource consumption, and unstable restoration. We propose a robust and efficient Diffusion-based Low-Light image enhancement approach, dubbed DiffLL.
  arXiv  Detail & Related papers  (2023-06-01T03:08:28Z)
• RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment [32.752633250862694]
  Generative foundation models are susceptible to implicit biases that can arise from extensive unsupervised training data. We introduce a new framework, Reward rAnked FineTuning, designed to align generative models effectively.
  arXiv  Detail & Related papers  (2023-04-13T18:22:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.